ACLU: Congress Cuts Medicaid to Fund ICE: How H.R. 1 Harms Communities

Congress Cuts Medicaid to Fund ICE: How H.R. 1 Harms Communities

Earlier this summer, Congress approved its most harmful budget in a generation. H.R. 1 or the “One Big Beautiful Bill Act” made the biggest cut to Medicaid since it was created in the 1960s and funneled that money to fund President Donald Trump’s racist anti-immigration agenda.

Instead of strengthening Medicaid, Congress took an axe to it. Instead of reining in Immigration and Customs Enforcement (ICE) abuses, Congress gave the agency billions more to terrorize our communities and detain families.

As Congress returns to their home states to meet with constituents and conduct in-district business during the August recess, the preceding six months must be top of mind. Representatives are facing the judgement of their constituents and must answer for their actions: did they capitulate to the Trump administration, or did they fight back? In turn, the American people must make clear that lawmakers' attacks on our health care and civil liberties are unacceptable.

Below, we outline what you need to know about this year's devastating federal budget bill — and how we can fight against this attack.

Cuts to Medicaid Threaten Our Lives

Medicaid is a critical resource across the country, especially for children and people with disabilities. Politicians' vote to cut Medicaid, a program that provides coverage for an estimated 70 million patients, means that in every single state and congressional district, people are closer to rationing medication, missing essential medical treatments, and losing access to the care they need.

Unsurprisingly, the most vulnerable among us will feel the brunt, including 12 million people with disabilities who rely on Medicaid. Medicaid coverage is a lynchpin to ensure people can exercise their right to live in their own homes, rather than dehumanizing institutions. Denying these supports and care forces people into institutions, stripping them of the liberty and autonomy our Constitution protects.

Martha Haythorn, a 25-year-old woman with Down syndrome from Georgia, is among the people whose freedom is endangered by cuts to Medicaid. She shared with the PBS Newshour that Medicaid helps her access her community. “I deserve to be there,” she said. “Without these benefits, I can't do that. Is it really worth taking away someone's benefit, someone's life, someone's accommodation?”

Courtney Leader of Missouri told CNN that she wrote to her senator, Josh Hawley, to share how Medicaid has helped her keep caring for her daughter, who has brain damage and cerebral palsy, at home. “Without Medicaid, we would lose everything – our home, our vehicles and, eventually, our daughter,” she wrote. Sen. Hawley was one of the 50 Senators who voted to cut Medicaid.

Medicaid also covers care for the two-thirds of people living in nursing facilities, and for the nearly 14 million people with mental health conditions or substance use disorders. Millions more could lose insurance coverage because of unclear rules and red tape created by new paperwork requirements and related reporting systems for Medicaid coverage.

The impact of H.R. 1 will be felt by people across the country, regardless of their incomes, for decades. Medicaid and other federal funds limited in H.R. 1 keep rural hospitals open, support medical training and clinical research, and ensure that people can get the care they need throughout their lifetimes. Finally, Congress’s cuts reduce the federal share of Medicaid funding, forcing states to cut services to meet budget shortfalls and further compounding the harm to our communities.

Access to health care is essential to our bodily autonomy and our freedom. Without these essential programs, our right to live fulfilled lives and control our futures will be greatly diminished.

Additional Risk to Reproductive Freedom

H.R. 1 also carries out one of the most dangerous goals of Project 2025: “defunding” Planned Parenthood by banning Medicaid patients from using their insurance there. This ban not only means that Medicaid patients can’t use their insurance for birth control, cancer screenings, STI testing, and other preventive care at Planned Parenthood — it also puts access to abortion access at risk even in states where abortion is legal.

A federal court blocked this “defunding” so Medicaid patients can continue to use their insurance for care at Planned Parenthood for now, but the legal battle will surely continue. If Planned Parenthood is truly “defunded”, hundreds of health centers would close, including a quarter of the country’s remaining abortion providers. Patients already struggle to access abortion, and, without Planned Parenthood, that could get even worse.

The ability to access timely reproductive health care will also worsen. There simply are not enough health care providers to serve the millions of patients seen by Planned Parenthood. The Trump administration has already attacked other essential reproductive health providers by withholding funds in the Title X family planning program, leaving patients with even fewer options for care. As a result, patients face longer wait times, travel, and delays in care.

For patients like Jamie Benner-Clemons, "defunding” Planned Parenthood could mean the difference between life and death. Benner-Clemons went to Planned Parenthood when she did not have health insurance. The cancer screening she received caught her breast cancer before it was too late. “The urgency of the health center staff saved my life,” she shared with Planned Parenthood.

Cuts to Medicaid will further endanger pregnant people nationwide. Medicaid covers 40 percent of births in the U.S. Without this coverage during pregnancy, many women would not get essential prenatal care. In rural areas, especially, these cuts will make pregnancy more dangerous. Dr. John Cullen, a family physician in Valdez, Alaska, a remote city of about 4,000 people, told The 19th, that “already we’re seeing [pregnancy care] deserts that are increasing in size, and after the passage of this bill those are going to be markedly worse — where people are going to have to drive hundreds of miles before they can get prenatal care, much less delivery.”

Cutting Health Care to Fund Detention Camps and Deportations

The politicians who supported H.R. 1 took health care away from millions, while giving $170 billion in taxpayer dollars to an immigration police force larger than most of the world’s militaries and detention camps that could, over time, hold 750,000 children, immigrants with legal status, and other long-time residents from our communities. This sends a clear message that, for politicians, care for people with disabilities is out, while $50,000 bonuses for new ICE agents are in.

This budget will reshape, perhaps permanently, immigration enforcement and detention in the U.S., funneling $45 billion into private prison companies and mass tent cities where few people can find a lawyer. This administration has already cut funding for services for unaccompanied children.

At a time when there are increased deaths in custody and appalling conditions at some of the new immigration camps, ICE is already cashing in on its new allowance and is set to expand detention to almost 110,000 beds in the next six months. This is dangerous and unprecedented. We need members of Congress to bring oversight to these expanding sites.

Meanwhile, in multiplying the number of immigration agents on our streets, H.R. 1 is funding a police state where parents—including U.S. citizens— who take their kids to school or the hospital may be asked to “show their papers” by masked agents or even be detained in private prisons and tent cities without counsel or due process.

We’ve already seen how the Trump administration’s immigration machine is terrorizing our communities. Masked ICE agents across the country have refused to provide identification during arrests, where they have forcibly taken people away in unmarked vehicles and left families with no information on where their loved ones are or whose custody they’ve been taken into. Already, there have been numerous cases of people impersonating ICE officers to threaten and harm people who are immigrants. As these raids and arrests grow more common and more aggressive, it is essential that immigration agents can be identified as such, which is why the VISIBLE Act — a commonsense transparency measure that requires DHS and ICE agents to wear visible identification — was just introduced in the Senate. As ICE agents multiply around the country with this new budget windfall, Congress must pass this bill immediately to protect community safety.

Americans Will Hold Congress Accountable

The politicians who voted for this budget monstrosity chose to sacrifice their constituents’ health, rights, and dignity. The American people will hold them accountable.

This August, the ACLU and our partners hosted events across the country to call out the politicians who supported this reckless attack on our health care, our civil liberties, and our very ability to survive. ACLU People Power volunteers mobilized their communities to town halls and confronted their Congress members about their decisions.

In Colorado, 50 constituents walked into Representative Jeff Hurd’s district lobby, holding signs and leaving hand-written notes at his office.

In Arizona, constituents attended a town hall hosted by the ACLU of Arizona in Representative David Schweikert’s district. Speakers and constituents alike focused on the $170 billion turbocharge of the president’s deportation machine, which has been fraught with abuses of power by the federal government and was paid for through cuts to Medicaid for people with disabilities.

Recent public polling demonstrates how extremely unpopular this bill is with Americans across the country. Two-thirds of the American public (64 percent) opposes H.R. 1. We must make this opposition clear to Congress by showing up at public events, calling, writing, and visiting congressional offices to let politicians know how unhappy constituents are, and how much H.R. 1 hurts our communities, our rights, and our society.

Republican leadership in Congress are preparing to create a multi-billion dollar slush fund for the Trump administration to use as their budget, as well as attach countless policy riders that risk our civil rights and liberties. We have the opportunity to tell our representatives to vote no on any of these poison pill spending bills that restrict our rights and we must. Join People Power today to stay informed about what Congress is up to and how you can get involved to protect all of our rights. Together we will fight and win.

Published August 20, 2025 at 11:51PM
via ACLU https://ift.tt/Vr57Svn

ACLU: Congress Cuts Medicaid to Fund ICE: How H.R. 1 Harms Communities

Congress Cuts Medicaid to Fund ICE: How H.R. 1 Harms Communities

Earlier this summer, Congress approved its most harmful budget in a generation. H.R. 1 or the “One Big Beautiful Bill Act” made the biggest cut to Medicaid since it was created in the 1960s and funneled that money to fund President Donald Trump’s racist anti-immigration agenda.

Instead of strengthening Medicaid, Congress took an axe to it. Instead of reining in Immigration and Customs Enforcement (ICE) abuses, Congress gave the agency billions more to terrorize our communities and detain families.

As Congress returns to their home states to meet with constituents and conduct in-district business during the August recess, the preceding six months must be top of mind. Representatives are facing the judgement of their constituents and must answer for their actions: did they capitulate to the Trump administration, or did they fight back? In turn, the American people must make clear that lawmakers' attacks on our health care and civil liberties are unacceptable.

Below, we outline what you need to know about this year's devastating federal budget bill — and how we can fight against this attack.

Cuts to Medicaid Threaten Our Lives

Medicaid is a critical resource across the country, especially for children and people with disabilities. Politicians' vote to cut Medicaid, a program that provides coverage for an estimated 70 million patients, means that in every single state and congressional district, people are closer to rationing medication, missing essential medical treatments, and losing access to the care they need.

Unsurprisingly, the most vulnerable among us will feel the brunt, including 12 million people with disabilities who rely on Medicaid. Medicaid coverage is a lynchpin to ensure people can exercise their right to live in their own homes, rather than dehumanizing institutions. Denying these supports and care forces people into institutions, stripping them of the liberty and autonomy our Constitution protects.

Martha Haythorn, a 25-year-old woman with Down syndrome from Georgia, is among the people whose freedom is endangered by cuts to Medicaid. She shared with the PBS Newshour that Medicaid helps her access her community. “I deserve to be there,” she said. “Without these benefits, I can't do that. Is it really worth taking away someone's benefit, someone's life, someone's accommodation?”

Courtney Leader of Missouri told CNN that she wrote to her senator, Josh Hawley, to share how Medicaid has helped her keep caring for her daughter, who has brain damage and cerebral palsy, at home. “Without Medicaid, we would lose everything – our home, our vehicles and, eventually, our daughter,” she wrote. Sen. Hawley was one of the 50 Senators who voted to cut Medicaid.

Medicaid also covers care for the two-thirds of people living in nursing facilities, and for the nearly 14 million people with mental health conditions or substance use disorders. Millions more could lose insurance coverage because of unclear rules and red tape created by new paperwork requirements and related reporting systems for Medicaid coverage.

The impact of H.R. 1 will be felt by people across the country, regardless of their incomes, for decades. Medicaid and other federal funds limited in H.R. 1 keep rural hospitals open, support medical training and clinical research, and ensure that people can get the care they need throughout their lifetimes. Finally, Congress’s cuts reduce the federal share of Medicaid funding, forcing states to cut services to meet budget shortfalls and further compounding the harm to our communities.

Access to health care is essential to our bodily autonomy and our freedom. Without these essential programs, our right to live fulfilled lives and control our futures will be greatly diminished.

Additional Risk to Reproductive Freedom

H.R. 1 also carries out one of the most dangerous goals of Project 2025: “defunding” Planned Parenthood by banning Medicaid patients from using their insurance there. This ban not only means that Medicaid patients can’t use their insurance for birth control, cancer screenings, STI testing, and other preventive care at Planned Parenthood — it also puts access to abortion access at risk even in states where abortion is legal.

A federal court blocked this “defunding” so Medicaid patients can continue to use their insurance for care at Planned Parenthood for now, but the legal battle will surely continue. If Planned Parenthood is truly “defunded”, hundreds of health centers would close, including a quarter of the country’s remaining abortion providers. Patients already struggle to access abortion, and, without Planned Parenthood, that could get even worse.

The ability to access timely reproductive health care will also worsen. There simply are not enough health care providers to serve the millions of patients seen by Planned Parenthood. The Trump administration has already attacked other essential reproductive health providers by withholding funds in the Title X family planning program, leaving patients with even fewer options for care. As a result, patients face longer wait times, travel, and delays in care.

For patients like Jamie Benner-Clemons, "defunding” Planned Parenthood could mean the difference between life and death. Benner-Clemons went to Planned Parenthood when she did not have health insurance. The cancer screening she received caught her breast cancer before it was too late. “The urgency of the health center staff saved my life,” she shared with Planned Parenthood.

Cuts to Medicaid will further endanger pregnant people nationwide. Medicaid covers 40 percent of births in the U.S. Without this coverage during pregnancy, many women would not get essential prenatal care. In rural areas, especially, these cuts will make pregnancy more dangerous. Dr. John Cullen, a family physician in Valdez, Alaska, a remote city of about 4,000 people, told The 19th, that “already we’re seeing [pregnancy care] deserts that are increasing in size, and after the passage of this bill those are going to be markedly worse — where people are going to have to drive hundreds of miles before they can get prenatal care, much less delivery.”

Cutting Health Care to Fund Detention Camps and Deportations

The politicians who supported H.R. 1 took health care away from millions, while giving $170 billion in taxpayer dollars to an immigration police force larger than most of the world’s militaries and detention camps that could, over time, hold 750,000 children, immigrants with legal status, and other long-time residents from our communities. This sends a clear message that, for politicians, care for people with disabilities is out, while $50,000 bonuses for new ICE agents are in.

This budget will reshape, perhaps permanently, immigration enforcement and detention in the U.S., funneling $45 billion into private prison companies and mass tent cities where few people can find a lawyer. This administration has already cut funding for services for unaccompanied children.

At a time when there are increased deaths in custody and appalling conditions at some of the new immigration camps, ICE is already cashing in on its new allowance and is set to expand detention to almost 110,000 beds in the next six months. This is dangerous and unprecedented. We need members of Congress to bring oversight to these expanding sites.

Meanwhile, in multiplying the number of immigration agents on our streets, H.R. 1 is funding a police state where parents—including U.S. citizens— who take their kids to school or the hospital may be asked to “show their papers” by masked agents or even be detained in private prisons and tent cities without counsel or due process.

We’ve already seen how the Trump administration’s immigration machine is terrorizing our communities. Masked ICE agents across the country have refused to provide identification during arrests, where they have forcibly taken people away in unmarked vehicles and left families with no information on where their loved ones are or whose custody they’ve been taken into. Already, there have been numerous cases of people impersonating ICE officers to threaten and harm people who are immigrants. As these raids and arrests grow more common and more aggressive, it is essential that immigration agents can be identified as such, which is why the VISIBLE Act — a commonsense transparency measure that requires DHS and ICE agents to wear visible identification — was just introduced in the Senate. As ICE agents multiply around the country with this new budget windfall, Congress must pass this bill immediately to protect community safety.

Americans Will Hold Congress Accountable

The politicians who voted for this budget monstrosity chose to sacrifice their constituents’ health, rights, and dignity. The American people will hold them accountable.

This August, the ACLU and our partners hosted events across the country to call out the politicians who supported this reckless attack on our health care, our civil liberties, and our very ability to survive. ACLU People Power volunteers mobilized their communities to town halls and confronted their Congress members about their decisions.

In Colorado, 50 constituents walked into Representative Jeff Hurd’s district lobby, holding signs and leaving hand-written notes at his office.

In Arizona, constituents attended a town hall hosted by the ACLU of Arizona in Representative David Schweikert’s district. Speakers and constituents alike focused on the $170 billion turbocharge of the president’s deportation machine, which has been fraught with abuses of power by the federal government and was paid for through cuts to Medicaid for people with disabilities.

Recent public polling demonstrates how extremely unpopular this bill is with Americans across the country. Two-thirds of the American public (64 percent) opposes H.R. 1. We must make this opposition clear to Congress by showing up at public events, calling, writing, and visiting congressional offices to let politicians know how unhappy constituents are, and how much H.R. 1 hurts our communities, our rights, and our society.

Republican leadership in Congress are preparing to create a multi-billion dollar slush fund for the Trump administration to use as their budget, as well as attach countless policy riders that risk our civil rights and liberties. We have the opportunity to tell our representatives to vote no on any of these poison pill spending bills that restrict our rights and we must. Join People Power today to stay informed about what Congress is up to and how you can get involved to protect all of our rights. Together we will fight and win.

Published August 20, 2025 at 07:21PM
via ACLU https://ift.tt/kiolt6E

ACLU: Will Giant Companies Always Have a Monopoly on Top AI Models?

Will Giant Companies Always Have a Monopoly on Top AI Models?

In my post on large language models (LLMs) last week, I argued that the most important question about LLMs is not the outcome of a race with China or when AI will reach human-level intelligence, but whether this striking new technology will be accessible to and serve the interests of ordinary people, or whether it will end up centrally controlled by a small number of highly capitalized companies or governments.

One of the biggest factors in determining which future we face is the question: how easy it is to access the resources needed to build and train a cutting-edge or “frontier” LLM? Inspired by this question, I did a bit of a dive into the training process for LLMs to try to assess the outlook for free LLMs. To explain what I found, I’ll need to look at the basics of how models are trained, as best I understand this fast-moving science from reading what experts in the field are saying.

A turnaround
As I mentioned in the prior post, things started off poorly. In 2022 and 2023, around the time OpenAI’s ChatGPT was released to the public, it looked like there was a real likelihood that this powerful new technology would be like nuclear power: centralized, complex, highly capital-intensive, secretive, and subject to strict security controls. The AI “base model” that was the engine driving the ChatGPT application was called GPT-3, which was succeeded by GPT-4 in 2023. Training these models on the vast amounts of data available on the internet and elsewhere was an enormously expensive undertaking that few organizations could afford. The GPT-4 model reportedly cost over $100 million to train, including the acquisition of 25,000 five-figure computer graphics cards and a huge electricity bill.

And it looked as though this level of resources was just the start. The seemingly miraculous performance of ChatGPT was largely the product of simply taking past research and scaling it up. LLMs that worked very poorly suddenly worked much better simply by exponentially increasing the amount of computing power (“compute”) dedicated to their training. Even though the computations were relatively simple — involving predicting the next “token” (roughly, word or symbol) in a text — when those computations were repeated trillions of times, unexpectedly smart behaviors appeared through a process known as “emergence.”

Emergence
Emergence is a phenomenon in complexity science referring to the fact that large numbers of simple rules can upon repetition produce complex and surprising behaviors that don’t appear to be the predictable result of any characteristics evident in the rules. A classic example is the flocking behavior of birds. Computer programmers trying to recreate that behavior in a computer bird simulation could tie themselves in knots trying to manually direct V-formations and the beautiful merging and diverging of flocks, but it turns out that if each virtual bird is programmed with just a few simple rules (“don’t get too close to your neighbors, but steer toward their average heading and position”) a flock of simulated birds will behave in ways strikingly similar to the complex movements of real flocks. That complex behavior emerges out of the simple rules in ways that nobody could ever predict by looking at the rules.

There are controversies over the precise meaning of “emergence” and the role it plays, but overall it seems to me that there’s no question that when AI systems are scaled up, they become capable of doing things that are surprising to people and far beyond anything they’ve been explicitly programmed to do. And that power is not to be underestimated; AI progress may not soar the way boosters predict — but neither should the technology be reductively dismissed as a mere “word prediction machine” or the like.

The original GPT-1 base model, released in 2018, had 117 million parameters — numbers that represent the strength of associations between different tokens, akin to synapses in the model’s “brain.” The next model, GPT-2, had over 12 times as many, and GPT-3 had 116 times more than that, powering the chatbot that burst into fame as ChatGPT. GPT-4 had an estimated ten times more and performed far better. Simply programming these models to predict the next word in a sequence led to emergent behaviors that seemed surprisingly (and deceptively) intelligent and human in some respects, and soared far beyond what one might expect from a system that, at root, is trained to simply predict the next word in a text.

In 2022 the lesson seemed clear: exponentially scaling up these models was the secret to success. At the rate things were progressing, it seemed plausible to many that with just a few more exponential leaps we might reach human-level intelligence — what is commonly if vaguely referred to as “artificial general intelligence,” or AGI. This fueled a “Manhattan Project” conception of LLM research as a geopolitical “race” toward a definitive goal: a sudden, secret breakthrough in reaching human or superhuman artificial intelligence. The winner of this race would obtain not a nuclear weapon but some sort of AI equivalent that would provide new levers for permanent dominance in business, the military, and the world. The implications of such a conception are bad for freedom: that research efforts should be concentrated and secretive, while cooperation and openness are foolish.

But a funny thing happened on the way to AGI: the benefits from scaling up the base models appear to have reached a state of significantly diminishing returns. For several years after GPT-4, for example, the AI world was eagerly awaiting, and OpenAI eagerly promoting, GPT-5, but its release was repeatedly delayed until, in late February 2025 the company finally released GPT-4.5 (suspected by some to be an expectations-lowering rename of GPT-5). It was by most accounts not dramatically higher-performing. “GPT 4.5 cost about 100x the compute of GPT-4 to train,” one expert observed, but “it is only slightly better on normal user metrics. Scaling as a product differentiator died in 2024.” Indeed, OpenAI’s competitors experienced a similar leveling off of progress in base-model training. The release of a model called GPT-5 in early August only confirmed this trend.

This was true even before the Chinese company DeepSeek made its enormous splash in December 2024 and January 2025, releasing models that achieved much more power at far lower training cost than previous models had been able to achieve. Some compared it to a company offering $50 smartphones more powerful than the latest $1,000 iPhone. This roiled stock valuations and was viewed largely through the lens of US-China geopolitical rivalries. The real significance of the DeepSeek innovations, however, was that it both clarified and accelerated the declining plausibility of the Manhattan Project model of LLM research.

The stages of training
As research continues, however, it is controversial and unclear to what extent the big science character of LLM training will continue to fade. The training of base models, where the returns to scale have apparently leveled off, at least for now, is only one step in creating LLMs. Meanwhile, other steps in the creation of finished models are being scaled up and absorbing more resources.

Overall, there are four basic stages in the creation of a model today, and they vary in what kinds of resources they require.

  a. Data preparation
Access to data for training the base model is the first thing that any actor wanting to train an LLM will need. The data that will be used to train the model must be selected, gathered, and perhaps filtered. Typically this means enormous masses of raw textual data (what is often summarized as “the entire Internet,” though it can also include the texts of books, messages, social media posts, and other things that may not be online). Images and video are increasingly being used as well for so-called “multi-modal” models that aim to understand images as well as text.

It’s not just big companies that can access all this data; there are a number of open data sets that are available for anyone to use. The most prominent, perhaps, is the nonprofit organization Common Crawl’s dataset, which includes regular snapshots of the entire public web gathered since 2008. Other publicly available datasets include the texts of books, Wikipedia, computer code repositories, and research papers. Experts say the biggest LLM players like OpenAI, Google, and Anthropic, reportedly do their own web crawling instead of just using such databases, and may have more resources to curate and filter the vast oceans of data that are poured into LLM pre-training. It’s not clear how much of an edge such work gives them in the quality of an LLM end product.

One of the reasons that the benefits of scaling up base-model training may have levelled off is that the amount of training data used has not scaled up commensurately with the number of parameters in the newest models at the frontier of research, according to experts. The internet and other prominent data sources have all been tapped by all the most prominent LLMs. But big companies that have exclusive access to other sources of data may as a result have an important leg up. Elon Musk’s model Grok, for example, has been trained on data from Twitter/X that nobody else can access. Other big LLM companies like OpenAI don’t have their own social media networks, and independent scientists certainly don’t. Again it is unclear how much of an advantage access to that kind of proprietary data will prove to be over time.

  b. Pre-training
Pre-training is the first step in actually creating a general LLM “base model” (as opposed to smaller or more specialized models derived from it). It involves teaching the model the basics of language by running the vast datasets through thousands of powerful graphical processing unit (GPU) cards to teach it to predict the next token in a string of text, building in the process a map of associations between different words and concepts. The output of this training is a set of model weights — essentially, a large set of numbers (in the largest models, trillions) representing the strength of associations or “thickness of the lines” between different tokens (corresponding roughly to neurons in the model’s “brain”).

As we have seen, pre-training is yielding diminishing returns, but it remains the most expensive step that uses the vast majority of the compute involved in creating a model.

There’s another technical development that may reduce the barriers to entry to training models. Most base-model training has taken place in specialized server farms with expensive GPUs packed close together because of the need to transfer enormous amounts of data many times between GPUs and very quickly (an ability known as “interconnect”) at each training step. This was seen as so vital that U.S. export controls targeted at China didn’t attempt to restrict compute, but only interconnect, on the assumption that would hobble China’s ability to do AI. But advances are being made in distributed training that allow far-flung computers to accomplish the same tasks. One expert, Nathan Labenz, observed that distributed training of LLMs as good as recent frontier models “is the kind of thing now that a well-organized but distributed group could probably potentially patch together the resources to do.”

  c. Post-training
In the post-training stage, a base model’s abilities in a specific area are shaped and refined through a variety of techniques. Supervised fine-tuning (SFT) can be used to refine the model’s abilities by providing it with a cultivated set of examples in a specialized area — for example, if you want the model to write about finance, health care, or the law, you might fine-tune it with data from those specialties. Another technique is “reinforcement learning from human feedback” (RLHF), in which humans give a model’s outputs thumbs ups and downs to nudge it toward behaving in certain ways and not others. A model can also be trained with synthetic data, feedback from another model, internal model self-critiques, or in some areas from self-play in which a model competes against itself.

Post-training is becoming an increasingly significant part of model building today. Several years ago it was mostly focused on style and safety, but its applications have gotten much broader. Post-training techniques are now used to shape a base model into a variety of derivative models. It could not only be trained to answer questions in a helpful and engaging manner for a chatbot, but also to actively search and retrieve new information (for “retrieval-augmented generation,” or RAG), to carry out tasks (for agents), to include images (for language vision models), or to specialize in answering objective scientific questions (for reasoning models).

Experts say that growth in the importance of post-training will likely continue. As Nathan Lambert of the nonprofit Allen AI research institute put it, “it’s very logical that post training will be the next domain for scaling model compute and performance” — meaning that like post-training scaling while it lasted, ever-larger resources may be needed to stay on the post-training cutting edge. Lambert points out that post-training is “still far cheaper than pretraining,” but that “post-training costs have been growing rapidly” into the tens of millions of dollars.

Still, there are many organizations that can spend tens of millions of dollars, compared to the billions of dollars that many expected model training to eventually cost. In addition, experts Like Labenz say that the lower technical difficulty of post-training makes it accessible to many more parties than base model training. “One of the biggest developments has been the recent revelation that reinforcement learning, on top of at least sufficiently powerful base models, really works and actually can be a pretty simple setup that works remarkably well,” he says.

  d. Inference
Inference-time compute is the processing that takes place after a user has made a query. When ChatGPT was first released, and for some time afterwards, LLMs did very little of this kind of processing, but the trend has been toward much more. Partly that is a result of reasoning models, which are created in post-training by giving LLMs large numbers of problems where there are objective right and wrong answers, a technique called reinforcement learning from verifiable rewards (RLVR). When training on problems with objective answers, such as in coding, science and math, models can be trained quickly and in great depth without human participation — teaching themselves, essentially. The researchers at Deepseek discovered that one of the emergent (spontaneously emerging) behaviors produced by such training was chain-of-thought reasoning, in which the model explicitly “thinks” step-by-step about the query it has received, explaining its reasoning along the way, and backtracking if necessary, before arriving at the answer. Although this emerges from training on science and math arenas, it appears that it generalizes to other, more subjective domains as well, making the model do better at all kinds of queries, including such things as legal reasoning. Wherever deployed, chain-of-thought reasoning increases the inference-time compute and thus the cost of running a model.

Another development that adds to the costs of inference is a trend toward larger “context windows.” In many ways LLMs are like the protagonist of the movie Memento or a dementia patient who can access a lifetime’s worth of background memory and knowledge, but is unable to form new memories. In answering queries, LLMs always have at hand the world-knowledge that they gained during pre- and post-training, but this knowledge is frozen. In terms of new input — what they can keep in “mind” during a single conversation — they have rather small short-term memories.

The original ChatGPT is thought to have had a context window of 8,192 tokens. Increasing context windows is expensive because in preparing an answer, an LLM must compare every token in the window to every other token. That means that the amount of processing compute needed rises roughly by the square of the number of tokens in memory. Nevertheless, some models now have relatively enormous context windows, such as Llama 4 Scout with 10 million tokens. This is in part due to clever innovations that are making it possible to reduce the compute needed to work with a window that big. But it’s still expensive. And even as inference becomes more compute-intensive, there is a strong demand for faster inference times, which is desirable for those who are want to use them for coding or for real-time applications like audio chat and live translation.

When inference costs are low, creating an LLM is like creating a railroad — it involves enormous upfront capital costs to build it, but then once built, relatively low marginal costs to run it. To the extent that inference costs grow, that raises ongoing operating costs. Providers are increasingly vying to offer large context windows and fast inference speeds — competing demands that are fast becoming major vectors of competition among LLM providers and reward computing power, scale, and centralization.

A diffusion of training ability
Overall, LLM research is in many ways “spreading outward” compared to its initial Manhattan Project-like character, and becoming more broadly accessible, giving us reason to hope that — especially with active measures by policymakers — LLMs may just not become the latest center of growing corporate power over individuals. The picture is complicated and fast-changing, however. It’s hard to know what direction progress will come from and how accessible the data, compute, interconnect, and other resources required for such progress will be. But that is the thing to watch — and the thing that policymakers should be actively trying to affect.

In the next installment of this series, I’ll look at another crucial factor in determining the future democratic character of LLMs: the state of open source LLM research and models.

Published August 20, 2025 at 04:37PM
via ACLU https://ift.tt/K0zfBZe

ACLU: New Detention Camp at Fort Bliss Marks Dangerous Expansion of Militarized Immigration Enforcement

New Detention Camp at Fort Bliss Marks Dangerous Expansion of Militarized Immigration Enforcement

The Trump administration opened an immigration detention camp at Fort Bliss, a U.S. military base in Texas. It is slated to become the largest immigration detention site in the country.

Fort Bliss, in El Paso, will hold up to 5,000 people in the base’s new tent camp. The military base is home to 90,000 service members and their families. Opened at the height of the Texas summer, where temperatures regularly soar above 100 degrees and sandstorms are frequent, detainees are at serious risk of heat-related illness and other harsh conditions. In response, local elected officials in El Paso have already passed a resolution demanding transparency and accountability from the federal government.

Built behind the walls of a military installation and away from public view, the facility is a calculated move to militarize immigration enforcement, reduce transparency, and fast-track deportations with minimal accountability.

Opening an immigration detention camp at Fort Bliss is just the latest step in President Donald Trump’s dystopian plan to detain and deport millions. In the coming months, the administration plans to open detention camps on at least two more military bases, with additional troops being deployed to support Immigrations and Customs Enforcement (ICE) operations nationwide. This coordinated use of the military to carry out a deportation agenda is unprecedented, and deeply dangerous. It’s a tactic straight out of the authoritarian playbook, and we cannot let it go unchallenged.

A Disturbing History of Fort Bliss as a Detention Site

Throughout history, Fort Bliss has served not as a place of refuge or justice, but as a staging ground for policies driven by cruelty and exclusion. Reopening it as an immigration detention site in 2025 would add yet another dark chapter to that legacy. Allowing this base to become the blueprint for exploiting military spaces and resources for immigration enforcement would set a dangerous precedent.

Fort Bliss has long been used to carry out government policies rooted in xenophobia and racism. During World War II, it was one of several Texas military bases used as internment camps for people of Japanese descent, as well as German and Italian immigrants. During the Mexican Revolution more than 20 years earlier, the base served as a temporary holding site for thousands of Mexican refugees, many of whom were kept in military tents and exposed to toxic chemical delousing agents. Some of the chemicals were later used in Nazi gas chambers.

In more recent history, Fort Bliss was used during Trump’s first term to detain children who were separated from their families. In 2021, under the Biden administration, government investigators reported children detained at the base experienced panic attacks and other forms of emotional distress due to inadequate staffing and improper care.

Funding Inhumane Detention

The Department of Homeland Security collaborating with the Department of Defense to detain thousands of people in tents on a military base is not only inhumane, it is ineffective and unnecessary. People held in remote facilities like Fort Bliss face steep barriers to accessing legal counsel and medical care, as well communicating with family members. Isolating people on military bases heightens the risk of abuse and neglect. It’s a punitive tactic designed to break people down, strip away hope, and pressure them into giving up their cases and accepting deportation.

This latest detention expansion is fueled by a massive influx of federal funding. Congress recently passed a reconciliation bill that allocated $170 billion toward immigration enforcement, including $45 billion specifically for detention. ICE’s detention budget is now 62 percent larger than the of the entire federal prison system. The Trump administration has already spent a staggering $1.2 billion on the Fort Bliss tent camp, rather than funding public services Texans urgently need, such as disaster relief, health care, education, and housing.

Texas already leads the nation in immigration detention, and the human cost is mounting. Since October, 13 people have died in ICE custody, already surpassing last year’s total. Meanwhile, ICE continues to obstruct oversight efforts by denying members of Congress and their staff access to detention facilities across the country, making transparency nearly impossible.

We Must Act Now to Stop This

El Paso County’s introduction of a resolution opposing the Fort Bliss detention center marks a critical first step in holding the federal government accountable. The ACLU applauds its call for transparency, local oversight, and the humane treatment of detainees. The public deserves to know what is happening behind the walls of its own military base. No one should be detained at Fort Bliss in the first place, and every individual held there deserves dignity, legal representation, and medical care.

In addition to state and local efforts, members of Congress must build on this momentum. They must demand full transparency from ICE by:

• Conducting oversight visits and releasing public reports on detention conditions.
• Opposing any more funding for militarized detention sites and tent cities and redeployment of military resources for the reckless deportation drive.
• Holding private contractors accountable for abuse and neglect.
• Supporting legislation that shifts immigration enforcement away from punitive approaches and toward permanent protections for immigrant communities.

The meaning of the word “Texas” comes from the Caddo word meaning “friend.” But this detention camp is anything but friendly. It is a cruel, reckless, and costly operation launched with little input from local communities or elected leaders. This is the Trump administration’s test case to see whether it can quietly convert military resources into tools for mass detention and deportation. We urge Congress and local leaders to act swiftly, decisively, and publicly. Our military bases are meant to defend our freedoms, not become the site of new human rights violations.

Published August 18, 2025 at 05:44PM
via ACLU https://ift.tt/UF4QO8K

ACLU: Live Coverage: ACLU Challenges ICE Detention Centers

Live Coverage: ACLU Challenges ICE Detention Centers


Published August 18, 2025 at 04:54PM
via ACLU https://ift.tt/FWAxu5y

ACLU: Live Coverage: ACLU Challenges ICE Detention Centers

Live Coverage: ACLU Challenges ICE Detention Centers


Published August 18, 2025 at 09:24PM
via ACLU https://ift.tt/TuLilQZ

ACLU: Florida's Secretive Immigration Detention Center, Explained

Florida's Secretive Immigration Detention Center, Explained

Deep in the Florida Everglades, a new immigration detention facility has appeared almost overnight. It’s the latest example of the Trump administration’s extreme approach to immigration enforcement, and it’s drawing significant, well-deserved outrage.

In just eight days, a former airstrip was converted into a large-scale immigration detention facility that now holds more than 700 detainees. Florida Attorney General James Uthmeier dubbed it “Alligator Alcatraz” to evoke the extreme isolation and cruel conditions of the now shut down federal prison -- leveraging the Everglades' vibrant wildlife as a potentially lethal barrier to human being

While Florida Governor Ron DeSantis and the Trump administration describe it as a temporary solution, the harsh, inhumane conditions and limited access to legal counsel pose a serious threat to human rights and due process. It is all part of Donald Trump’s broader policy of mass detention and deportation that disregards legal protections and basic human dignity.

Below, we explain what you need to know about the cruelly dubbed “Alligator Alcatraz,” and what it reveals about the Trump administration’s escalating approach to immigration enforcement.

The Camp Was Built in Secret and Without Proper Approval

The state of Florida never held a public hearing, obtained full local approval, or went through an environmental review process before seizing the county-owned land used for this facility. Governor DeSantis said he was acting in response to a federal request and authorized the project under emergency powers granted by a 2023 executive order. Using these powers, he bypassed laws intended to safeguard public transparency and environmental protection, citing an alleged immigration “emergency” as justification.

Local officials in Collier County, where the facility is located, were reportedly unaware of the site until days before its opening. This group includes the county’s planning and zoning director, emergency management director, and a commissioner, who only learned about it when DeSantis, Trump, and federal officials toured the site.

This facility sets a troubling precedent by pushing the boundaries of state authority in immigration enforcement. It comes at a time when Congress has recently increased funding by billions for ICE detention and deportation efforts, and when the federal government holds primary responsibility for immigration enforcement.

Conditions Inside the Camp Are Dangerous and Inhumane

The first group of detainees arrived at the facility on July 3rd, and in the short time since, hundreds are now detained in makeshift tents and cages, exposed to extreme heat, mosquitoes, and unsanitary conditions, and surrounded by snakes and alligators. Congressional visitors reported cages holding 30 or more individuals at temperatures near 100 degrees. Detainees report minimal and sometimes contaminated food, bug infestations, overflowing toilets, and lack of medical care — conditions so severe that they have led the detainees to stage a hunger strike.

Although the facility is promoted as having capacity for up to 3,000 detainees — with Department of Homeland Security (DHS) officials claiming it could eventually hold 5,000 — it lacks permanent infrastructure. The facility depends on temporary Federal Emergency Management Agency (FEMA) trailers, portable air conditioning units, and mobile power and water systems. While government officials have described it as a “low-cost” facility, operating expenses are projected at $450 million annually, initially paid by Florida taxpayers and later reimbursed by federal agencies. Even officials involved in planning the site acknowledge that the harsh conditions were deliberately designed to encourage undocumented people who are already in America to self-deport.

People Are Being Denied Legal Access and Due Process

Multiple lawyers reported being turned away by armed National Guard and state police officers at the entrance gate. Detainees have no reliable way to make confidential calls to lawyers. Documents shared between attorneys and clients are being reviewed by staff — a direct violation of attorney-client privilege.

DeSantis has requested that National Guard members serve as immigration judges at the facility, expediting deportations and stripping away independent review. The Florida Governor describes this as cutting through “bureaucracy,” but in reality, it eliminates fair hearings and replaces impartial judges with political appointees.

These practices violate constitutional rights guaranteed by the First and Fifth Amendments, not only for detainees but for their attorneys. In response, on July 16, the American Civil Liberties Union, the ACLU of Florida, and Americans for Immigrant Justice filed a lawsuit,C.M. v. Noem, challenging these violations on behalf of detainees and legal service providers.

This Facility is Part of a Larger Expansion of Detention and Deportation

The Everglades detention center is a key example of the Trump administration’s coordinated expansion of detention and deportation nationwide. As of late June, more than 58,000 immigrants were in ICE custody, many held in local jails because ICE only has funding to house an average of 41,000 people. ICE is increasingly relying on private contractors — including for security and facility management — to run these operations and has been sending people who are immigrants to El Salvador’s CECOT prison and unconventional detention facilities like military bases.

DeSantis has called the facility, which allows the government to fly people who are immigrants directly out from the old airstrip, a “force multiplier” for the administration’s mass deportation effort. Despite the harmful narrative surrounding immigrants and crime, only a third of people at these facilities have been charged for minor offenses, such as illegal reentry or traffic violations.

The Landscape Puts Lives at Risk — and the Facility Could Do Permanent Damage to the Environment

The Everglades detention center was built in the middle of a fragile, flood-prone wetland ecosystem, threatening local wildlife and increasing the risk of long-term environmental damage. Trump officials have claimed the swamp — with its alligators and snakes — adds to the site’s “security.” But there’s nothing secure about housing thousands of people in tents and trailers during hurricane season, especially with only vague plans for evacuation if disaster strikes.

First responders in Collier County have raised concerns about which agency would be responsible if an emergency occurred at the site. The county’s emergency management director, Dan Summers, was not even informed about the facility’s construction plans. Undocumented people are being housed in repurposed FEMA trailers and tents, the same types often used to shelter people displaced by natural disasters like hurricanes. These structures will do little to protect people from what forecasters predict may be an active hurricane season.

The detention center doesn’t only put detainees at risk — it could also affect nearby communities, water systems, and the broader region. The Everglades are a protected wetland and a crucial source of drinking water for millions of South Floridians. Placing thousands of people so close to this vital water source raises serious concerns about contamination. Environmental groups, including Friends of the Everglades, the Center for Biological Diversity, and Earthjustice sued the DeSantis administration on June 27 to stop the project and its harmful impact on the surrounding Big Cypress National Preserve. Not only is the area a key water supply for Floridians, but it is also home to endangered species like the Florida panther, and the disruption caused by the detention center could push them closer to extinction. The facility also violates the rights of the Miccosukee Tribe of Indians of Florida, whose sacred lands border the airport and runway. The tribe has joined the lawsuit.

Where the Fight Goes from Here

The Trump administration cannot sidestep constitutional protections by relocating detention centers to remote locations or calling them “temporary.” We can expect further expansion and the spread of ICE tactics given the influx of federal funding approved by Congress this summer, including greater involvement of state and local agencies as they seek to broaden their reach across the country.

The ACLU remains committed to challenging the use of this facility in court, defending the rights of those held without due process, legal access, or safe conditions. Detaining individuals under cruel and chaotic circumstances is not immigration policy — it is abuse. We will continue to fight until these unlawful practices end.

Published August 15, 2025 at 11:57PM
via ACLU https://ift.tt/oEAeCp1

ACLU: Florida's Secretive Immigration Detention Center, Explained

Florida's Secretive Immigration Detention Center, Explained

Deep in the Florida Everglades, a new immigration detention facility has appeared almost overnight. It’s the latest example of the Trump administration’s extreme approach to immigration enforcement, and it’s drawing significant, well-deserved outrage.

In just eight days, a former airstrip was converted into a large-scale immigration detention facility that now holds more than 700 detainees. Florida Attorney General James Uthmeier dubbed it “Alligator Alcatraz” to evoke the extreme isolation and cruel conditions of the now shut down federal prison -- leveraging the Everglades' vibrant wildlife as a potentially lethal barrier to human being

While Florida Governor Ron DeSantis and the Trump administration describe it as a temporary solution, the harsh, inhumane conditions and limited access to legal counsel pose a serious threat to human rights and due process. It is all part of Donald Trump’s broader policy of mass detention and deportation that disregards legal protections and basic human dignity.

Below, we explain what you need to know about the cruelly dubbed “Alligator Alcatraz,” and what it reveals about the Trump administration’s escalating approach to immigration enforcement.

The Camp Was Built in Secret and Without Proper Approval

The state of Florida never held a public hearing, obtained full local approval, or went through an environmental review process before seizing the county-owned land used for this facility. Governor DeSantis said he was acting in response to a federal request and authorized the project under emergency powers granted by a 2023 executive order. Using these powers, he bypassed laws intended to safeguard public transparency and environmental protection, citing an alleged immigration “emergency” as justification.

Local officials in Collier County, where the facility is located, were reportedly unaware of the site until days before its opening. This group includes the county’s planning and zoning director, emergency management director, and a commissioner, who only learned about it when DeSantis, Trump, and federal officials toured the site.

This facility sets a troubling precedent by pushing the boundaries of state authority in immigration enforcement. It comes at a time when Congress has recently increased funding by billions for ICE detention and deportation efforts, and when the federal government holds primary responsibility for immigration enforcement.

Conditions Inside the Camp Are Dangerous and Inhumane

The first group of detainees arrived at the facility on July 3rd, and in the short time since, hundreds are now detained in makeshift tents and cages, exposed to extreme heat, mosquitoes, and unsanitary conditions, and surrounded by snakes and alligators. Congressional visitors reported cages holding 30 or more individuals at temperatures near 100 degrees. Detainees report minimal and sometimes contaminated food, bug infestations, overflowing toilets, and lack of medical care — conditions so severe that they have led the detainees to stage a hunger strike.

Although the facility is promoted as having capacity for up to 3,000 detainees — with Department of Homeland Security (DHS) officials claiming it could eventually hold 5,000 — it lacks permanent infrastructure. The facility depends on temporary Federal Emergency Management Agency (FEMA) trailers, portable air conditioning units, and mobile power and water systems. While government officials have described it as a “low-cost” facility, operating expenses are projected at $450 million annually, initially paid by Florida taxpayers and later reimbursed by federal agencies. Even officials involved in planning the site acknowledge that the harsh conditions were deliberately designed to encourage undocumented people who are already in America to self-deport.

People Are Being Denied Legal Access and Due Process

Multiple lawyers reported being turned away by armed National Guard and state police officers at the entrance gate. Detainees have no reliable way to make confidential calls to lawyers. Documents shared between attorneys and clients are being reviewed by staff — a direct violation of attorney-client privilege.

DeSantis has requested that National Guard members serve as immigration judges at the facility, expediting deportations and stripping away independent review. The Florida Governor describes this as cutting through “bureaucracy,” but in reality, it eliminates fair hearings and replaces impartial judges with political appointees.

These practices violate constitutional rights guaranteed by the First and Fifth Amendments, not only for detainees but for their attorneys. In response, on July 16, the American Civil Liberties Union, the ACLU of Florida, and Americans for Immigrant Justice filed a lawsuit,C.M. v. Noem, challenging these violations on behalf of detainees and legal service providers.

This Facility is Part of a Larger Expansion of Detention and Deportation

The Everglades detention center is a key example of the Trump administration’s coordinated expansion of detention and deportation nationwide. As of late June, more than 58,000 immigrants were in ICE custody, many held in local jails because ICE only has funding to house an average of 41,000 people. ICE is increasingly relying on private contractors — including for security and facility management — to run these operations and has been sending people who are immigrants to El Salvador’s CECOT prison and unconventional detention facilities like military bases.

DeSantis has called the facility, which allows the government to fly people who are immigrants directly out from the old airstrip, a “force multiplier” for the administration’s mass deportation effort. Despite the harmful narrative surrounding immigrants and crime, only a third of people at these facilities have been charged for minor offenses, such as illegal reentry or traffic violations.

The Landscape Puts Lives at Risk — and the Facility Could Do Permanent Damage to the Environment

The Everglades detention center was built in the middle of a fragile, flood-prone wetland ecosystem, threatening local wildlife and increasing the risk of long-term environmental damage. Trump officials have claimed the swamp — with its alligators and snakes — adds to the site’s “security.” But there’s nothing secure about housing thousands of people in tents and trailers during hurricane season, especially with only vague plans for evacuation if disaster strikes.

First responders in Collier County have raised concerns about which agency would be responsible if an emergency occurred at the site. The county’s emergency management director, Dan Summers, was not even informed about the facility’s construction plans. Undocumented people are being housed in repurposed FEMA trailers and tents, the same types often used to shelter people displaced by natural disasters like hurricanes. These structures will do little to protect people from what forecasters predict may be an active hurricane season.

The detention center doesn’t only put detainees at risk — it could also affect nearby communities, water systems, and the broader region. The Everglades are a protected wetland and a crucial source of drinking water for millions of South Floridians. Placing thousands of people so close to this vital water source raises serious concerns about contamination. Environmental groups, including Friends of the Everglades, the Center for Biological Diversity, and Earthjustice sued the DeSantis administration on June 27 to stop the project and its harmful impact on the surrounding Big Cypress National Preserve. Not only is the area a key water supply for Floridians, but it is also home to endangered species like the Florida panther, and the disruption caused by the detention center could push them closer to extinction. The facility also violates the rights of the Miccosukee Tribe of Indians of Florida, whose sacred lands border the airport and runway. The tribe has joined the lawsuit.

Where the Fight Goes from Here

The Trump administration cannot sidestep constitutional protections by relocating detention centers to remote locations or calling them “temporary.” We can expect further expansion and the spread of ICE tactics given the influx of federal funding approved by Congress this summer, including greater involvement of state and local agencies as they seek to broaden their reach across the country.

The ACLU remains committed to challenging the use of this facility in court, defending the rights of those held without due process, legal access, or safe conditions. Detaining individuals under cruel and chaotic circumstances is not immigration policy — it is abuse. We will continue to fight until these unlawful practices end.

Published August 15, 2025 at 07:27PM
via ACLU https://ift.tt/m3iXNEw

ACLU: Many Are Focused on the Wrong Questions When it Comes to AI

Many Are Focused on the Wrong Questions When it Comes to AI

Discussing the history of weapons, George Orwell argued that some, like tanks, naturally lend themselves to despotism because they are complex, expensive and difficult to make, while others, like muskets and rifles, are “inherently democratic.” I’ve always remembered that notion, so when ChatGPT burst into public consciousness in November of 2022, I immediately thought about how much ChatGPT and other Large Language Models (LLMs) looked like a tank.

Since then, the technology has evolved in dramatic and often surprising ways, and today the situation is much less clear — there are now reasons to hope that LLMs will end up less like tanks and more like muskets. In 2022 LLMs were a technology resting in the hands of the few giant tech companies with the expertise, access to data, and deep pockets to create it. There were already various stripes of “AI,” but LLMs seemed more powerful and more centralized than anything that had been seen before. Today there are more players able to build models, a general commodification of models, and a plausible path toward open-source LLMs that can compete at the top tier. Models have also become far more compressed, so that useful work can be done locally rather than through cloud servers controlled by a few big companies.

The technology world has been following the evolution of LLMs with rapt attention — and rightly so. But many people have been asking the wrong questions: Who will win a geopolitical battle for AI dominance, the US or China? Will the technology evolve into human-level “Artificial General Intelligence” (AGI)? When will LLMs start becoming effective “agents” not just processing information but helping people perform tasks?

Some of these are interesting questions, but none are as important for the freedom and empowerment of ordinary people as the question: Who will AI empower? That is a far more urgent question than the stock valuation of big tech companies, speculative musings about AGI, or the state of a US-China race for dominance. The shape of LLM science has crucial implications for intellectual freedom, scientific research, the democratization of information, control over access to technology, privacy, and ultimately democracy itself.

Beyond Orwell’s “tank vs. musket” framing, the notion that technologies carry an inherent politics has also been explored by thinkers such as Langdon Winner, who looked at how their qualities can reflect and reinforce specific power structures. Nuclear power, he argues, inherently requires hierarchical management, rigorous security regimes, and centralized control because of its scale, complexity, and danger. Solar power, on the other hand, is more compatible with the values of decentralization and democracy because it can be implemented at small scales, requires less specialized expertise, and doesn’t pose catastrophic risks that require intense security measures.

Longstanding questions about AI — and new ones
Beyond the built-in characteristics of a technology, of course, a lot depends on the particularities of design and deployment. For example, the copy machine, which was used by Soviet dissidents, Daniel Ellsberg, and others to distribute forbidden information, might be a more inherently democratic technology than the broadcast television station, which naturally lends itself to despotic government. But the effects of centralized broadcast television can be neutralized through careful protections such as independent control, free speech rights, and the guarding of diversity and competition. (Before the internet, those were very prominent public issues in the United States to which groups like the ACLU devoted a lot of attention.) Conversely, the potentially pro-freedom tendencies of other technologies can be neutralized — for example when copy machines include fingerprinting technology that can link printouts to particular machines and even operators. Solar power may lend itself to decentralized deployment, but it could certainly be implemented in a centralized, authoritarian manner.

The various forms of AI have long raised civil liberties and fairness issues, especially those around transparency, the composition of training data, bias, automated decision making, inappropriate deployments, and due process. Many of those issues have remained substantially the same whether the algorithm is a neural network trained on millions of data points or a formula in a spreadsheet.

But the advent of LLMs (based on an entirely different technology, known as transformers, than most previous AI products) has intensified and expanded those issues — and raised new ones:

• LLMs intensify transparency concerns because they operate in an even more opaque and unpredictable manner than other AI systems.
• The data on which they are trained can be even harder to evaluate.
• The models’ appearance of greater intelligence is likely tempting more people to use them in more decisionmaking roles, even as their biases and other irrationalities remain at least as strong as in other forms of AI.
• LLMs appear to be supercharging the use of AI in communications and video

The same policy battles that have been fought over automated algorithms for years will continue to be fought over LLMs. But LLMs also increase the stakes of all of the above because they are beginning to play an educational and linguistic role that no other form of AI has approached — potentially influencing how people write, communicate, and potentially think across a wide variety of contexts.

The stakes
Aside from longstanding issues around the deployment of AI, it is important to consider Orwell’s question — whether the technology itself is shaping up to be inherently biased toward democracy or authoritarianism. And that will depend both on policy choices that we make and on how the science develops. On one hand, it’s possible that the technology has had its quantum evolutionary leap and is now stalling out — that it is and will remain a “normal technology,” even if a gradually transformative one such as electricity or the Internet. On the other hand, if the technology does improve rapidly, including conceivably to some form of human-like conscious intelligence, then the power profile of this technology will matter all the more.

The worst case scenario is that we end up in a world where one or a handful of entities control the LLMs that (for whatever reason) everyone uses. If that happens, those players will come to wield enormous power, even if the technology improves only gradually or progress stalls out.

• Those who control this tool will gain the dangerous power to filter the acquisition of knowledge as people increasingly use these tools for research about the world, analysis of their own data, the production of writing and media, and perhaps as agents that perform tasks besides fetching information.
• They will be able to do that by choosing which deep-seated biases to try to correct and which to ignore, and perhaps by instantiating certain biases into the products in the first place — biases that may be subtle and very hard to detect or measure. Think of a more subtle version of the “Great Firewall of China,” which the government uses to engage in mass filtering and censorship of what information people in China can access. Large companies are temperamentally conservative and do not generally support significant challenges to the status quo of which they are a significant part. And historically, big companies have accommodated themselves to authoritarian governments.
• An LLM monopoly or oligarchy is also likely to try to keep any secrets about advances and improvements in the technology to themselves, stifling democratic access and scientific inquiry. And they could have the power to exclude some parties from using their LLMs at all, much like the credit card oligopoly today, which blocks payments by sexually oriented businesses and journalists disfavored by government officials.
• They’re also likely to surveil everyone who uses the models. From the moment that ChatGPT arrived on the scene, there has been speculation that LLMs would displace search and become an enormous source of advertising revenue. The data that can be collected ranges from people’s LLM queries — like search, an enormous source of sensitive data — to the documents and videos they upload for analysis, to the text of personal therapeutic or “friend” chats.

In short, LLMs could provide the newest form of dangerously concentrated power, following in the footsteps of the big tech giants of today.

In better-case scenarios, on the other hand, LLMs could empower individuals in positive ways. Rather than remaining in the hands of a few, a thousand flowers could bloom as a flourishing marketplace of diverse models trained for all kinds of specialties emerges, many of them transparent and open source and small enough to run on local computers under the control of individuals. Just as the printing press broke the medieval Catholic Church’s near-monopoly on the ability to read, interpret, and publish the written word, LLMs might democratize skills and abilities that are currently held by only a relatively small elite, such as the ability to program computers and create apps. They could allow reporters or citizens to search and analyze overwhelming volumes of government or corporate data for reporting and oversight, put the ability to create a feature-length film in the hands of anyone, and democratize many other things that are now the exclusive domain of experts or well-funded businesses.

What to watch
So how are we to evaluate whether LLMs are tilting in democratic or authoritarian directions? The technology is developing rapidly and unpredictably, and since the advent of ChatGPT there have been dramatic developments that bear directly on the balance between the above outcomes. In particular, there are three interrelated areas that have significant implications for freedom:

• The degree to which training and running an LLM emerges as a form of big science — large-scale, high-cost projects like the Manhattan Project, physics supercolliders, or space telescopes — or whether training the models people want to use ends up being a broadly accessible thing.
• The ability to run desirable models on local hardware, which will remove the possibility of AI titans engaging in gatekeeping, censorship, and privacy invasion.
• The health of open source models and research, which will help ensure that no company enjoys a monopoly on models people want to use.

I will take a closer look at these areas in follow-up posts.

Published August 14, 2025 at 06:44PM
via ACLU https://ift.tt/u970LQh

ACLU: Many Are Focused on the Wrong Questions When it Comes to AI

Many Are Focused on the Wrong Questions When it Comes to AI

Discussing the history of weapons, George Orwell argued that some, like tanks, naturally lend themselves to despotism because they are complex, expensive and difficult to make, while others, like muskets and rifles, are “inherently democratic.” I’ve always remembered that notion, so when ChatGPT burst into public consciousness in November of 2022, I immediately thought about how much ChatGPT and other Large Language Models (LLMs) looked like a tank.

Since then, the technology has evolved in dramatic and often surprising ways, and today the situation is much less clear — there are now reasons to hope that LLMs will end up less like tanks and more like muskets. In 2022 LLMs were a technology resting in the hands of the few giant tech companies with the expertise, access to data, and deep pockets to create it. There were already various stripes of “AI,” but LLMs seemed more powerful and more centralized than anything that had been seen before. Today there are more players able to build models, a general commodification of models, and a plausible path toward open-source LLMs that can compete at the top tier. Models have also become far more compressed, so that useful work can be done locally rather than through cloud servers controlled by a few big companies.

The technology world has been following the evolution of LLMs with rapt attention — and rightly so. But many people have been asking the wrong questions: Who will win a geopolitical battle for AI dominance, the US or China? Will the technology evolve into human-level “Artificial General Intelligence” (AGI)? When will LLMs start becoming effective “agents” not just processing information but helping people perform tasks?

Some of these are interesting questions, but none are as important for the freedom and empowerment of ordinary people as the question: Who will AI empower? That is a far more urgent question than the stock valuation of big tech companies, speculative musings about AGI, or the state of a US-China race for dominance. The shape of LLM science has crucial implications for intellectual freedom, scientific research, the democratization of information, control over access to technology, privacy, and ultimately democracy itself.

Beyond Orwell’s “tank vs. musket” framing, the notion that technologies carry an inherent politics has also been explored by thinkers such as Langdon Winner, who looked at how their qualities can reflect and reinforce specific power structures. Nuclear power, he argues, inherently requires hierarchical management, rigorous security regimes, and centralized control because of its scale, complexity, and danger. Solar power, on the other hand, is more compatible with the values of decentralization and democracy because it can be implemented at small scales, requires less specialized expertise, and doesn’t pose catastrophic risks that require intense security measures.

Longstanding questions about AI — and new ones
Beyond the built-in characteristics of a technology, of course, a lot depends on the particularities of design and deployment. For example, the copy machine, which was used by Soviet dissidents, Daniel Ellsberg, and others to distribute forbidden information, might be a more inherently democratic technology than the broadcast television station, which naturally lends itself to despotic government. But the effects of centralized broadcast television can be neutralized through careful protections such as independent control, free speech rights, and the guarding of diversity and competition. (Before the internet, those were very prominent public issues in the United States to which groups like the ACLU devoted a lot of attention.) Conversely, the potentially pro-freedom tendencies of other technologies can be neutralized — for example when copy machines include fingerprinting technology that can link printouts to particular machines and even operators. Solar power may lend itself to decentralized deployment, but it could certainly be implemented in a centralized, authoritarian manner.

The various forms of AI have long raised civil liberties and fairness issues, especially those around transparency, the composition of training data, bias, automated decision making, inappropriate deployments, and due process. Many of those issues have remained substantially the same whether the algorithm is a neural network trained on millions of data points or a formula in a spreadsheet.

But the advent of LLMs (based on an entirely different technology, known as transformers, than most previous AI products) has intensified and expanded those issues — and raised new ones:

• LLMs intensify transparency concerns because they operate in an even more opaque and unpredictable manner than other AI systems.
• The data on which they are trained can be even harder to evaluate.
• The models’ appearance of greater intelligence is likely tempting more people to use them in more decisionmaking roles, even as their biases and other irrationalities remain at least as strong as in other forms of AI.
• LLMs appear to be supercharging the use of AI in communications and video

The same policy battles that have been fought over automated algorithms for years will continue to be fought over LLMs. But LLMs also increase the stakes of all of the above because they are beginning to play an educational and linguistic role that no other form of AI has approached — potentially influencing how people write, communicate, and potentially think across a wide variety of contexts.

The stakes
Aside from longstanding issues around the deployment of AI, it is important to consider Orwell’s question — whether the technology itself is shaping up to be inherently biased toward democracy or authoritarianism. And that will depend both on policy choices that we make and on how the science develops. On one hand, it’s possible that the technology has had its quantum evolutionary leap and is now stalling out — that it is and will remain a “normal technology,” even if a gradually transformative one such as electricity or the Internet. On the other hand, if the technology does improve rapidly, including conceivably to some form of human-like conscious intelligence, then the power profile of this technology will matter all the more.

The worst case scenario is that we end up in a world where one or a handful of entities control the LLMs that (for whatever reason) everyone uses. If that happens, those players will come to wield enormous power, even if the technology improves only gradually or progress stalls out.

• Those who control this tool will gain the dangerous power to filter the acquisition of knowledge as people increasingly use these tools for research about the world, analysis of their own data, the production of writing and media, and perhaps as agents that perform tasks besides fetching information.
• They will be able to do that by choosing which deep-seated biases to try to correct and which to ignore, and perhaps by instantiating certain biases into the products in the first place — biases that may be subtle and very hard to detect or measure. Think of a more subtle version of the “Great Firewall of China,” which the government uses to engage in mass filtering and censorship of what information people in China can access. Large companies are temperamentally conservative and do not generally support significant challenges to the status quo of which they are a significant part. And historically, big companies have accommodated themselves to authoritarian governments.
• An LLM monopoly or oligarchy is also likely to try to keep any secrets about advances and improvements in the technology to themselves, stifling democratic access and scientific inquiry. And they could have the power to exclude some parties from using their LLMs at all, much like the credit card oligopoly today, which blocks payments by sexually oriented businesses and journalists disfavored by government officials.
• They’re also likely to surveil everyone who uses the models. From the moment that ChatGPT arrived on the scene, there has been speculation that LLMs would displace search and become an enormous source of advertising revenue. The data that can be collected ranges from people’s LLM queries — like search, an enormous source of sensitive data — to the documents and videos they upload for analysis, to the text of personal therapeutic or “friend” chats.

In short, LLMs could provide the newest form of dangerously concentrated power, following in the footsteps of the big tech giants of today.

In better-case scenarios, on the other hand, LLMs could empower individuals in positive ways. Rather than remaining in the hands of a few, a thousand flowers could bloom as a flourishing marketplace of diverse models trained for all kinds of specialties emerges, many of them transparent and open source and small enough to run on local computers under the control of individuals. Just as the printing press broke the medieval Catholic Church’s near-monopoly on the ability to read, interpret, and publish the written word, LLMs might democratize skills and abilities that are currently held by only a relatively small elite, such as the ability to program computers and create apps. They could allow reporters or citizens to search and analyze overwhelming volumes of government or corporate data for reporting and oversight, put the ability to create a feature-length film in the hands of anyone, and democratize many other things that are now the exclusive domain of experts or well-funded businesses.

What to watch
So how are we to evaluate whether LLMs are tilting in democratic or authoritarian directions? The technology is developing rapidly and unpredictably, and since the advent of ChatGPT there have been dramatic developments that bear directly on the balance between the above outcomes. In particular, there are three interrelated areas that have significant implications for freedom:

• The degree to which training and running an LLM emerges as a form of big science — large-scale, high-cost projects like the Manhattan Project, physics supercolliders, or space telescopes — or whether training the models people want to use ends up being a broadly accessible thing.
• The ability to run desirable models on local hardware, which will remove the possibility of AI titans engaging in gatekeeping, censorship, and privacy invasion.
• The health of open source models and research, which will help ensure that no company enjoys a monopoly on models people want to use.

I will take a closer look at these areas in follow-up posts.

Published August 14, 2025 at 11:14PM
via ACLU https://ift.tt/xQWuMAP