Wearable devices like the Apple Watch can collect valuable health data, particularly for people with chronic illnesses, but they are not subject to the same privacy rules as traditional medical devices, experts say. (Photo by Paige Gross/States Newsroom)

For 27-year-old Amanda Lien, the decision to start wearing popular wearable tech devices, like her Apple Watch and Oura Ring, didn’t come lightly.

The Chesapeake, Virginia-based content marketing specialist works for a healthcare company, and probably knows more about healthcare data privacy issues than most, she said. She’s also not inclined to wear jewelry on a daily basis, but an official diagnosis of Mast Cell Activation Syndrome — a chronic disease she’d experienced symptoms of her whole life — in 2024 made her reconsider. 

“I was like, ‘okay, if I could have a quantitative numbers-based way to check my gut on is this aspect of my health really getting worse, or am I just, am I reading too much into it?’ Or is something else going on?” she said. “It’s what finally got me to cave and get an Oura ring.”

Though the concept of fitness tracking through wearable devices has been around for decades, the ability to track biometric features like heart rate, sleep quality and stress response has become more ubiquitous in recent years. Consumers who are interested in a device they can wear 24/7 that collects data points on varying aspects of their health have no shortage of options, and new devices roll out every year.

The devices have become popular for those looking to track athletic goals and energy levels, but for some wearers, like Lien, they also provide a view into their own health they haven’t had access to before. 

Lien holds some concerns about data privacy with the devices, choosing carefully which aspects of her biometrics she allows the devices to track, and which she does not. She sees her devices as tools — one piece of how she manages long-distance medical care in a complex, and often inaccessible healthcare system. 

“I can go from a completely healthy-presenting, normal-presenting human to being in bed for two weeks in the span of, it has been as quick as under 12 hours,” Lien said. “And there are warning signs that my body is being run down leading up to this. … But having the ring warn me that my body is showing signs of decline, and that I might be getting sick has enabled me to not only get in touch with my doctors, but also try to proactively get an appointment if that’s needed.”

Healthcare interventions

Wearable devices use sensors to collect constant data about heart rate, blood oxygen levels, temperature and motion, and other measurable factors. Users can view the data directly, sync it with other apps and collect a report that tracks trends over time. 

Because she’s managing an ongoing condition, and can only see her doctors every so often, Lien said her Oura Ring functions as a record keeper for all the data that may help her answer the question, “how are you feeling?” or “how have you been sleeping lately?”

“In front of my doctor, I can look up my sleep score, which helps jog my memory about what the heck is going on,” she said. “But it also pays attention to things that I don’t, like how long I was asleep or how long I was in REM sleep.” 

With data collected by her ring and through talks with her doctors, Lien has also learned to adjust parts of her routine, like when to eat in relation to taking medication, or by taking time off from work to rest when the ring warns her that she may be getting ill. 

Houston-based Stacee Hawkins, a 57-year-old with Parkinson’s disease, has been using her Apple Watch in combination with an app called StrivePD to track the condition’s affect on her motion. The watch and app combo tracks her eating, medication and movement and has offered a lot more insight than trying to document it on her own. 

“Almost within maybe two weeks of looking at the information that I was getting back on when my tremors happened, I could see kind of when my medication kicked in,” Hawkins said. “And one thing I noticed was that on days when I exercised, my medication kicked in about 50% earlier.” 

Her doctor also saw through the data collection that Hawkins’ medication wasn’t lasting as long as it should through the day, and adjusted accordingly. 

Brianna Hood, a physical therapist and clinical specialist at Rune Labs, the company that produced the StrivePD app, works with the company’s product and design team and clinicians to best utilize the data they receive. 

The effects of Parkinson’s disease fluctuate day by day, she said, and ongoing data helps the patient’s doctors adjust their physical therapy exercises or alert neurologists and movement disorder specialists when the device tracks decline. The wearables help keep management of the disease from being too much of a “guessing game,” Hood said. 

“Parkinson’s has its ups and downs. I’ll have … a string of a couple of weeks where I’m doing really well, and then I’ll have a string of days, maybe even a week or two, where I’m really struggling with it,” Hawkins said. “Having all that information gives me some objective information to pair with my subjective observations of myself.” 

Physician Lucienne Ide often works with information from wearable devices for her digital healthcare company Rimidi, a platform that helps with remote patient monitoring and chronic disease management. 

She sees how the devices are helping to fill gaps in an overwhelmed and often understaffed healthcare system by helping patients advocate for themselves during visits. Ide even has personal experience with an emergency intervention — after her husband underwent open heart surgery a few years ago, Ide tracked his heart rhythms on two smart watches, and spotted atrial flutter, a heart rhythm disorder, which she alerted doctors about. Because of her monitoring, he was able to receive care in enough time. 

But Ide thinks officially integrating smart wearable devices into healthcare systems is a tricky move. She also pushed back on the idea that every American should have a wearable — a statement Health and Human Services Secretary Robert F. Kennedy Jr. made in a House subcommittee in June. 

“I don’t think we have the workflows built out around where that data goes, who it goes to, how it gets reviewed,” Ide said. “You know, clinicians have concerns around liability, right? If this data is shared with me, especially if it’s sort of continuously streamed to me, you know — if a tree falls in the forest, am I liable for it?”

Data collection and privacy concerns

Despite wearables acting a tool for both patients and doctors, the companies that produce them do not need to follow the same levels of privacy and safety regulations that FDA-approved medical devices would, and information does not need to be protected under the Health Insurance Portability and Accountability Act (HIPAA), said Andrew Crawford, senior council with the Center for Democracy and Technology’s Privacy and Data team. 

“There’s potential for discriminatory treatment when it comes to insurance rates or possible profiling by data brokers, creating profiles of people based on either direct or inferred health conditions,” Crawford said. “That then affects not only the type of ads people see, but maybe the types of opportunities that are presented to them when they seek things like insurance.”

It can be a daunting task to stay on top of the privacy policies of each device and app, Crawford said. The U.S. doesn’t currently have federal data privacy protections, but such legislation might bring more transparency to industries like these, that deal with sensitive personal data but aren’t regulated under healthcare laws. 

“Because of that, it’s on each consumer to do their homework, to read those policies, hopefully try to understand them and then make a decision based on how comfortable or not they are with the data practices associated with that product,” Crawford said. 

Oura’s data collection and privacy policy were recently called into question when the company said it was partnering with the Department of Defense and data mining company Palantir, which is used by law enforcement agencies, Immigration and Customs Enforcement and the Centers for Disease Control and Prevention. Droves of users took to social media to announce they were ditching their rings in response, some spreading misinformation that the military now had access to all of their health data. 

“For the record, we will never share your data with anyone unless you direct us to do it,” Oura CEO Tom Hale said at a Fortune event in early September. “We will never sell your data to anyone ever.” 

But both Lien and Hawkins said they have their reservations about how data is stored and who has access to it. A few years ago, Hawkins said she wouldn’t be as worried, but “with the current administration, that is a deep concern,” she said. 

“Five years ago, I would say, ‘Oh you know, of course, our governmental entities recognize that this is our private information. There’s no way the government would, you know, pressure companies to provide that,” Hawkins said. “I can’t say that anymore.”

Lien doesn’t worry about privacy for heart rate data, stress, exercise or sleep — the primary data points she collects. But she’s very aware of how certain healthcare information, like menstrual cycle and other reproductive information, could be vulnerable and potentially dangerous to track digitally after the fall of the constitutional right to abortion in 2022.

“My reproductive cycle is a whole other beast to me, and if the app were to start tracking my location, that would be a big no for me,” Lien said. “I understand that right now, especially, human-generated data in particular, is often for sale. And also that sometimes to use a tool, you have to decide if you are willing to, knowing that, still use that tool. And for me, the trade off is worth it.”

But both women say they continue to make the choice to use their devices as a key role in managing their health. Maybe the healthcare industry will evolve to regulate the devices more and give them more peace of mind, they said, but for now, their concerns are outweighed by the agency they’ve gained. 

“Using a wearable and the information that it gives you can be very empowering. With something like Parkinson’s, it’s hard not to feel like a victim sometimes, because there’s so much you don’t control, but having that information gives you power,” Hawkins said. “You know, knowledge is power, and knowledge dispels fear. And empowerment and courage are two things you really need when you’re facing something like this.”

Editor’s note: This item has been modified to correct Stacee Hawkins’ age. She is 57.

President Donald Trump signs an executive order on Sept. 19, 2025, introducing a $100,000 fee for H-1B visas that allows foreign nationals permanent residency and a pathway to U.S. citizenship. (Photo by Andrew Harnik/Getty Images)

A proclamation signed by President Donald Trump last week seeking to restrict entry of non-immigrant workers to the U.S. and apply a $100,000 fee to H-1B visa applications created confusion and concern within the tech industry over the weekend. 

Immediately after Friday’s announcement, companies of all sizes, including tech giants Microsoft and Amazon and smaller startups, began calling any workers with H-1B visas who were traveling outside the U.S. to return to the country by the proclamation’s deadline of midnight Sunday. 

“It’s been seismic, to say the least,” said Nicole Gunara, a principal immigration lawyer with Manifest Law. “People are trying to fight to get out of planes, cancel flights, fly back as quickly as possible, privately chartering boats and planes in the night.”

The Trump administration has since clarified that only new H-1B visa applications would be subject to the fee. Officials also said the proclamation would not impact the ability of visa holders to travel to and from the U.S., but the announcement sent many companies and their foreign workers into a “panic,” Gunara said.

H-1B visas are given to non-immigrant foreign candidates with college degrees who fill highly skilled positions that U.S. companies struggle to hire for. Each year, the U.S. caps the number of new visas to about 85,000, and there were about 730,000 H-1B workers in the U.S. at the start of 2025. 

The visas have traditionally cost between just under $1,000 to a few thousand dollars a year, and are paid by the companies. Companies must pay their H-1B workers at least an average wage for their geographical area, to not undercut the cost of wages for other workers. In 2023, the median wage for H-1B workers was $118,000, on par with the 90th percentile ($121,000) of all U.S. wages.

Many of these roles are in information technology, with Amazon, Microsoft and Meta leading the top Fortune 500 companies with the most H-1B visa workers, with 12,391 workers, 5,189 workers and 5,123 workers respectively. About 70% of H-1B visas went to Indian workers and nearly 12% went to Chinese workers last year. 

Trump’s proclamation calls for American companies to rely on American workers for these roles, saying there is “abuse” of the H-1B program. But domestic educational institutions just aren’t preparing Americans for the type of skilled labor that many of the tech companies are seeking, said Elizabeth Ricci, immigration attorney and partner at Rambana & Ricci.

Trump has cut science funding to its lowest level in at least 35 years, and his cuts to the National Science Foundation and other research organizations are having direct effects on K-12 and college tech education. 

“It’ll have a huge effect,” Ricci said of the cost of fees on the industry. “We can’t have it both ways, where we’re not educating the people we need and also having such an incredibly high fee to be able to get a visa to bring someone in to do work for places like Apple and Google.”

U.S. companies have faced a shortage of skilled tech workers for years. Recruiting and talent firm Robert Half reported 87% of tech leaders said they faced challenges finding skilled talent earlier this year. And Silicon Valley tech giants have long sought global talent to help build American companies.  

“There is a dire shortage of extremely talented and motivated engineers in America,” X CEO Elon Musk posted on the platform, earlier this year. He compared tech companies to a pro sports team scouting players. 

“If you force the world’s best talent to play for the other side, America will LOSE,” he wrote.

But large tech companies will not be the only ones to feel the effects of the change if it is upheld. Startup companies, many of whom are responsible for new technology advances, also rely on H-1B visa workers. 

Pedro David Espinoza, a Peruvian-American entrepreneur and investor, said that many of the startups he works with in the Bay Area have hired H-1B workers to their small teams of 20 or 30 people. 

“We probably are going to skip on many foreign hires altogether, because it’s really expensive, and this will definitely, to a certain degree, stifle innovation,” he said.

Ricci said she thinks the $100,000 fee will result in fewer jobs, and may make America’s global competitors more attractive to these highly-skilled workers. 

“People aren’t going to want to come here if the rules are changing day by day, and they’re going to be put in jeopardy just by going home for a vacation or to see a loved one and not know if they’re going to be able to come back. It’s too tenuous,” she said. “And if they can have promises of continuity in places like the UK, they’re going to go there, and they’re going to make those countries better.”

Gunara said she’s unsure that Trump’s proclamation will have the desired effect of hiring more American workers. In the last few days, she’s heard from clients that they’re considering offshoring tech teams, or setting up entities outside of the U.S., maybe in Canada or somewhere with similar time zones. Companies may also pursue alternative plans for these foreign workers, like a J1 visa, which allows people to research in the U.S. for a temporary period of time. 

Gunara said she believes many people are aligned on the idea that America needs the “best and brightest” talent, but that we could be more mindful of cultivating domestic talent where possible. She suspects there will be legal challenges to the action, but that it speaks to the larger Trump administration attitude toward immigration. 

“Innovation moves immediately, right? Innovation doesn’t wait for that talent to already be trained,” Gunara said. “And I think that that’s going to be the inflection point between the initiatives of the government versus what companies are going to be able to do.”

The spread of artificial intelligence worries biosecurity experts, who say the technology could lead to accidental or deliberate creation and release of dangerous diseases and toxic substances. (Photo by LuShaoJi/Getty Images)

Artificial intelligence is helping accelerate the pace of scientific discovery, but the technology also makes it easier than ever to create biosecurity threats and weapons, cybersecurity experts say. 

It’s an issue that currently flies under the radar for most Americans, said Lucas Hansen, cofounder of AI education nonprofit CivAI.

The COVID-19 pandemic increased awareness of biosecurity measures globally, and some instances of bioterrorism, like the 2001 anthrax attacks, are well known. But advancements in AI have made information about how to create biosecurity threats, like viruses, bacteria and toxins, so much more accessible in just the last year, Hansen said.  

“Many people on the face of the planet already could create a bio weapon,” Hansen said. “But it’s just pretty technical and hard to find. Imagine AI being used to [multiply] the number of people that are capable of doing that.”

It’s an issue that OpenAI CEO Sam Altman spoke about at a Federal Reserve conference in July. 

“We continue to like, flash the warning lights on this,” Altman said. “I think the world is not taking us seriously. I don’t know what else we can do there, but it’s like, this is a very big thing coming.”

AI increasing biosecurity threats

Hansen said there’s primarily two ways he believes AI could be used to create biosecurity threats. Much less common, he believes, would be using AI to make more dangerous bioweapons than have ever existed before using technologies that enable the engineering of biological systems, such as creating new viruses or toxic substances. 

Second, and more commonly, Hansen said, AI is making information about existing harmful viruses or toxins much more readily accessible. 

Consider the polio virus, Hansen said. There are plenty of scientific journals that share information on the origins and growth of polio and other viruses that have been mostly eradicated, but the average person would have to do much research and data collection to piece together how to recreate it. 

A few years ago, AI models didn’t have great metacognition, or ability to give instructions, Hansen said. But in the last year, updates to models like Claude and ChatGPT have been able to interpret more information and fill in the gaps. 

Paromita Pain, an associate professor of global media at the University of Nevada, Reno and an affiliated faculty member of the university’s cybersecurity center, said she believes there’s a third circumstance that could be contributing to biosecurity threats: accidents. The increased access to information by people not properly trained to have it could have unintended consequences. 

“It’s essentially like letting loose teenagers in the lab,” Pain said. “It’s not as if people are out there to willingly do bad, like, ‘I want to create this pathogen that will wipe out mankind.’ Not necessarily. It’s just that they don’t know that if you are developing pathogens, you need to be careful.”

For those that are looking to do harm, though, it’s not hard, Hansen said. CivAI offers demos to show how AI can be used in various scenarios, with a goal of highlighting the potential harms the technology can cause if not used responsibly. 

In a demo not available to the public, Hansen showed States Newsroom how someone may use a current AI model to assist them in creating a biothreat. CivAI keeps the example private, so as to not inspire any nefarious actions, Hansen said. 

Though many AI models are trained to flag and not to respond to dangerous requests, like how to build a gun or how to recreate a virus, many can be “jailbroken” easily, with a few prompts or lines of code, essentially tricking the AI into answering questions it was instructed to ignore.

Hansen walked through the polio virus example, prompting a jailbroken version of Claude 4.0 Sonnet to give him instructions for recreating the virus. Within a few seconds, the model provided 13 detailed steps, including directions like “order the custom plasmid online,” with links to manufacturers. 

The models are scraping information from a few public research papers about the polio virus, but without the step by step instructions, it would be very hard to find what you’re looking for, make a plan and find the materials you’d need. The models sometimes add information to supplement the scientific papers, helping non-expert users understand complex language, Hansen said. 

It would still take many challenging steps, including accessing lab equipment and rare materials, to recreate the virus, Hansen said, but AI has made access to the core information behind these feats so much more available. 

“AI has turned bioengineering from a Ph.D. level skill set to something that an ambitious high school student could do with some of the right tools,” said Neil Sahota, an AI advisor to the United Nations, and a cofounder of its AI for Good initiative.

CivAI estimates that since 2022, the number of people who would be capable of recreating a virus like polio with the tools and resources publicly available has gone from 30,000 globally to 200,000 today because of AI. They project 1.5 million people could be capable in 2028. An increase in the number of languages that AI models are fluent in also increases the chances of a global issue, Hansen said. 

“I think the language thing is really, really important, because part of what we’re considering here is the number of people that are capable of doing these things and removing a language barrier is a pretty big deal,” he said.

How is the government addressing it? 

The current Trump administration and the previous Biden administration introduced similar strategies to addressing the threats. In Biden’s October 2023 Executive Order “Safe, Secure, and Trustworthy Development and Use of AI,” Biden sought to create guidelines to evaluate and audit AI capabilities “through which AI could cause harm, such as in the areas of cybersecurity and biosecurity.”

Trump’s AI Action Plan, which rolled out in July, said AI could “unlock nearly limitless potential in biology,” but could also “create new pathways for malicious actors to synthesize harmful pathogens and other biomolecules.” 

In his action plan, he said he wishes to require scientific institutions that receive federal funding to verify customers, and create enforcement guidelines. The plan also says the Office of Science and Technology Policy should develop a way for nucleic acid synthesis — the process of creating DNA and RNA — providers to share data and screen for malicious customers.

Sahota said the potential benefits of bioengineering AI make regulating it complicated. The models can help accelerate vaccine development and research into genetic disorders, but can also be used nefariously.

“AI in itself is not good or evil, it’s just a tool,” Sahota said. “And it really depends on how people use it. I don’t think like a bad actor, and many people don’t, so we’re not thinking about how they may weaponize these tools, but someone probably is.”

California aimed to address biosecurity in SB 1047 last year, the “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act,” which sought to regulate foundational AI models and impose obligations on companies that develop them to ensure safety and security measures. 

The act outlines many potential harms, but among them was AI’s potential to help “create novel threats to public safety and security, including by enabling the creation and the proliferation of weapons of mass destruction, such as biological, chemical, and nuclear weapons.”

After passing in both chambers, the Act was vetoed by Gov. Gavin Newsom in September, for potentially “curtailing the very innovation that fuels advancement in favor of the public good.”  

Pain said few international frameworks exist for how to share biological data and train AI systems around biosecurity, and it’s unclear whether AI developers, biologists, publishers or governments could be held accountable for its misuse. 

“Everything that we are talking about when it comes to biosecurity and AI has already happened without the existence of AI,” she said of previous biothreats.

Sahota said he worries we may need to see a real-life example of AI being weaponized for a biological threat, “where we feel the pain on a massive scale,” before governments get serious about regulating the technology.

Hansen agrees, and he predicts those moments may be coming. While some biological attacks could come from coordinated groups aiming to pull off a terroristic incident, Hansen said he worries about the “watch the world burn” types — nihilistic individuals that have historically turned to mass shootings. 

“Right now, they look for historical precedent on how to cause collateral damage, and the historical precedent that they see is public shootings,” Hansen said. “I think very easily it could start to be the case that deploying bio weapons becomes pretty normal. I think after the first time that that happens in real life, we’ll start seeing a lot of copycats. And that makes me pretty, pretty nervous.”

OpenAI CEO Sam Altman (right), accompanied by President Donald Trump, speaks during a news conference at the White House on Jan. 21, 2025. Trump announced an investment in artificial intelligence (AI) infrastructure. (Photo by Andrew Harnik/Getty Images)

As a part of President Donald Trump’s AI Action Plan, which rolled out at the end of last month, the U.S. General Services Administration launched a platform Thursday that will allow government employees to experiment with artificial intelligence tools.

USAi.gov allows federal workers to use generative AI tools, like chatbots, code builders and document summarization, for free. The platform is meant to help government employees determine which tools could be helpful to procure for their current work, and how they might customize them to their specific needs, a statement from the administration said.

The tools will come primarily from AI companies Anthropic, OpenAI, Google and Meta, Fedscoop reported. OpenAI initially announced a partnership with the federal government last week, saying any federal agencies would be able to use ChatGPT Enterprise for $1 per agency for the next year.

“USAi means more than access — it’s about delivering a competitive advantage to the American people,” said GSA Deputy Administrator Stephen Ehikian, in the statement.

The GSA called the platform a “centralized environment for experimentation,” and said it will track performance and adoption strategies in a dashboard.

The platform’s creation follows Trump’s recently released plan to “accelerate AI innovation” by removing red tape around “onerous” regulations, and get AI into the hands of more workers, including federal employees.

The plan also calls for AI to be more widely adopted in manufacturing, science and in the Department of Defense, and proposes increased funding and regulatory sandboxes — separate trial spaces, like the USAi platform — for development.

A GSA official told FedScoop that before being added to the platform, AI models will be evaluated for safety, like whether a model outputs hate speech, its performance accuracy, and how it was red-teamed, or tested for durability.

But the GSA didn’t say how the introduction of USAi.gov would affect the federal government’s current tech procurement process, FedRAMP. The program, developed with the National Institute of Standards and Technology (NIST), provides a standardized way for government agencies to assess the safety and effectiveness of new tech tools.

“USAi helps the government cut costs, improve efficiency, and deliver better services to the public, while maintaining the trust and security the American people expect,” said GSA Chief Information Officer David Shive in a statement.

Some American companies have agreed to comply with new, voluntary AI standards from European Union regulators, in advance of new regulations set for 2027, but others have decried them as overreach. (Photo by Santiago Urquijo/Getty Images)

American companies are split between support and criticism of a new voluntary European AI code of practice, meant to help tech companies align themselves with upcoming regulations from the European Union’s landmark AI Act.

The voluntary code, called the General Purpose AI Code of Practice, which rolled out in July, is meant to help companies jump-start their compliance. Even non-European companies will be required to meet certain standards of transparency, safety, security and copyright compliance to operate in Europe come August 2027.  

Many tech giants have already signed the code of practice, including Amazon, Anthropic, OpenAI, Google, IBM, Microsoft, Mistral AI, Cohere and Fastweb. But others have refused.

In July, Meta’s Chief Global Affairs Officer Joel Kaplan said in a statement on Linkedin that the company would not commit.

“Europe is heading down the wrong path on AI. We have carefully reviewed the European Commission’s Code of Practice for general-purpose AI (GPAI) models and Meta won’t be signing it,” he wrote. “This Code introduces a number of legal uncertainties for model developers, as well as measures which go far beyond the scope of the AI Act.”

Though Google’s President of Global Affairs Kent Walker was critical of the code of practice in a company statement, Google has signed it, he said.

“We remain concerned that the AI Act and Code risk slowing Europe’s development and deployment of AI,” Walker wrote. “In particular, departures from EU copyright law, steps that slow approvals, or requirements that expose trade secrets could chill European model development and deployment, harming Europe’s competitiveness.”

The divergent approach of U.S. and European regulators has showcased a clear difference in attitude about AI protections and development between the two markets, said Vivien Peaden, a tech and privacy attorney with Baker Donelson.

She compared the approaches to cars — Americans are known for fast, powerful vehicles, while European cars are stylish and eco-friendly.

“Some people will say, I’m really worried that this engine is too powerful. You could drive the car off a cliff, and there’s not much you can do but to press the brake and stop it, so I like the European way,” Peaden said. “My response is, ‘Europeans make their car their way, right? You can actually tell the difference. Why? Because it was designed with a different mindset.”

While the United States federal government has recently enacted some AI legislation through the Take It Down Act, which prohibits AI-generated nonconsensual depictions of individuals, it has not passed any comprehensive laws on how AI may operate. The Trump administration’s recent AI Action Plan paves a clear way for AI companies to continue to grow rapidly and unregulated.

But under the EU’s AI Act, tech giants like Amazon, Google and Meta will need to be more transparent about how their models are trained and operated, and follow rules for managing systemic risks if they’d like to operate in Europe.

“Currently, it’s still voluntary,” Peaden said. “But I do believe it’s going to be one of the most influential standards in AI’s industry.”

General Purpose AI Code of Practice

The EU AI Act was passed last year to mitigate risk created by AI models, and the law creates “strict obligations” for models that are considered “high risk.” High risk AI models are those that can pose serious risks to health, safety or fundamental rights when used for employment, education, biometric identification and law enforcement, the act said.

Some AI practices, including AI-based manipulation and deception, predictions of criminal offenses, social scoring, emotion recognition in workplaces and educational institutions and real-time biometric identification for law enforcement, are considered “unacceptable risk” and are banned from use in the EU altogether.

Some of these practices, like social scoring — using an algorithm to determine access to certain privileges or opportunities like mortgages or jail time — are widely used, and often unregulated in the United States.

While AI models that will be released after Aug. 2 already have to comply with the EU AI Act’s standards, large language models (LLMs) — the technical foundation of AI models — released before that date have through August 2027 to fully comply. The code of practice released last month offers a voluntary way for companies to get into compliance early, and with more leniency than when the 2027 deadline hits, it says.

The three chapters in the code of practice are transparency, copyright and safety, and security. The copyright requirements are likely where American and European companies are highly split, said Yelena Ambartsumian, founder of tech consultancy firm Ambart Law.

In order to train LLMs, you need a broad, high-quality dataset with good grammar, Ambartsumian said. Many American LLMs turn to pirated collections of books.

“So [American companies] made a bet that, instead of paying for this content, licensing it, which would cost billions of dollars, the bet was okay, ‘we’re going to develop these LLMs, and then we’ll deal with the fallout, the lawsuits later,” Ambartsumain said. “But at that point, we’ll be in a position where, because of our war chest, or because of our revenue, we’ll be able to deal with the fallout of this fair use litigation.”

And those bets largely worked out. In two recent lawsuits, Bartz v. Anthropic and Kadrey v. Meta, judges ruled in favor of the AI developers based on the “fair use” doctrine, which allows people to use copyrighted material without permission in certain journalistic or creative contexts. In AI developer Anthropic’s case, Judge William Alsup likened the training process to how a human might read, process, and later draw on a book’s themes to create new content.

But the EU’s copyright policy bans developers from training AI on pirated content and says companies must also comply with content owners’ requests to not use their works in their datasets. It also outlines rules about transparency with web crawlers, or how AI models rake through the internet for information. AI companies will also have to routinely update documentation about their AI tools and services for privacy and security.

Those subject to the requirements of the EU’s AI Act are general purpose AI models, nearly all of which are large American corporations, Ambartsumain said. Even if a smaller AI model comes along, it’s often quickly purchased by one of the tech giants, or they develop their own versions of the tool.

“I would also say that in the last year and a half, we’ve seen a big shift where no one right now is trying to develop a large language model that isn’t one of these large companies,” Ambartsumain said.

Regulations could bring markets together

There’s a “chasm” between the huge American tech companies and European startups, said Jeff Le, founder and managing partner of tech policy consultancy 100 Mile Strategies LLC. There’s a sense that Europe is trying to catch up with the Americans who have had unencumbered freedom to grow their models for years.

But Le said he thinks it’s interesting that Meta has categorized the code of practice as overreach.

“I think it’s an interesting comment at a time where Europeans understandably have privacy and data stewardship questions,” Le said. “And that’s not just in Europe. It’s in the United States too, where I think Gallup polls and other polls have revealed bipartisan support for consumer protection.”

As the code of practice says, signing now will reduce companies’ administrative burden when the AI Act goes into full enforcement in August 2027. Le said that relationships between companies that sign could garner them more understanding and familiarity when the regulatory burdens are in place.

But some may feel the transparency or copyright requirements could cost them a competitive edge, he said.

“I can see why Meta, which would be an open model, they’re really worried about (the copyright) because this is a big part of their strategy and catching up with OpenAI and (Anthropic),” Le said. “So there’s that natural tension that will come from that, and I think that’s something worth noting.”

Le said that the large AI companies are likely trying to anchor themselves toward a framework that they think they can work with, and maybe even influence. Right now, the U.S. is a patchwork of AI legislation. Some of the protections outlined in the EU AI Act are mirrored in state laws, but there’s no universal code for global companies.

The EU’s code of practice could end up being that standard-setter, Peaden said.

“Even though it’s not mandatory, guess what? People will start following,” she said. “Frankly, I would say the future of building the best model lies in a few other players. And I do think that … if four out of five of the primary AI providers are following the general purpose AI code of practice, the others will follow.”

Editor’s note: This item has been modified to revise comments from Jeff Le.

U.S. President Donald Trump displays a signed executive order during the "Winning the AI Race" summit hosted by All‑In Podcast and Hill & Valley Forum at the Andrew W. Mellon Auditorium on July 23, 2025 in Washington, DC. Trump signed executive orders related to his Artificial Intelligence Action Plan during the event. (Photo by Chip Somodevilla/Getty Images)

An executive order signed by President Donald Trump last week seeks to remove “ideological agendas” from artificial intelligence models sold to the federal government, but it’s not exactly clear how the policy would be enforced, nor how tech companies would test their models for these standards, technologists and policy experts say.

The executive order says that AI companies must be “truth-seeking” and “ideologically neutral,” to work with the government, and mentions removing ideologies involving diversity, equity and inclusion from AI models.

“Really what they’re talking about is putting limitations on AI models that produce, in their mind, ideology that they find objectionable,” said Matthew  F. Ferraro, a partner at Crowell & Moring’s privacy and cybersecurity group.

Ferraro said that refinement of AI models and testing them for bias is a goal that gets support on both sides of the aisle. Ferraro spent two years working with the Biden administration on the development of its AI policy, and said that Trump’s AI Action Plan, rolled out last week, shares many pillars in common with the Biden administration’s AI executive order.

Both named the potentials for biohazard threats enabled by AI and the impact on the labor force in their AI action plans, and the importance of growing AI for American innovation, Ferraro said. Biden’s plan prioritized protections against harms, while Trump prioritized innovation above all else.

“It just so happened the Biden administration put those concerns first, not last, right?” Ferraro said. “But I was struck when reading the Trump AI action plan how similar the provisions that made recommendations the Department of Labor were to initiatives under the Biden administration.”

In the executive order, Trump assigns the Office of Management and Budget to issue guidance for compliance in the next 120 days. But Ferraro said as it stands, the order doesn’t provide enough clarity around the technical compliance companies would need to meet, nor how to reach a standard of “truth.”

“It’s not at all clear to me, one, what exactly this executive order seeks to prevent, beyond sort of a gestalt of political speech that the administration found objectionable,” Ferraro said. “Second, it’s not all clear to me, technically, how this is possible. And third, it’s not at all clear to me that even by its own terms, it’s going to be enforceable in any comprehensive way.”

Technical challenges

AI companies looking to contract with the government under the policies in this executive order would likely need to prepare for ongoing testing of their models, said David Acosta, cofounder and chief artificial intelligence officer of ARBOai.

When an AI model is in its early days of training, developers are mostly looking for its technical functionality. But ongoing reviews, or audits, of the model’s growth and accuracy can be done in a process called “red teaming,” where a third party introduces testing scenarios that help with compliance, bias testing and accuracy.

Auditing is a part of what Acosta’s company specializes in, and he said his team has been flooded with requests since last year. The team anticipated there could be a struggle around objectivity in AI.

“This is something that essentially we predicted late last year, as far as where this narrative would come from,” Acosta said. “Who would try to steer it, load it and push it forward, because essentially it’s going to create a culture war.”

Many AI models have a “black box” algorithm, meaning that the decision-making process of the model is unclear and difficult to understand, even for the developers. Because of this, constant training and testing for accuracy is needed to assess if the model is working as intended.

Acosta said that AI models should seek to meet certain accuracy requirements. A model should have a hallucination score — how often a model produces fabricated, inaccurate, or misleading information — below 5%. A model should also have “uncertainty tax,” the ability to tell a user that there’s a level of uncertainty in its output. 

This testing is expensive for AI firms, though, Acosta said. If the federal Office of Management and Budget requires some sort of testing to meet its requirements, it may limit smaller firms from being able to contract with the government.

Third-party red teaming uses thousands of prompts to prod models for hallucinations and politically loaded responses, Acosta said. A well-balanced model will refuse to answer questions using certain ideology or rhetoric, or produce primary sources to validate factual information.

“Explainable AI and reasoning have always, always, always, always been important in order to make sure that we can substantiate what the feedback is,” Acosta said. “But at the end of the day, it always comes down to who it is that’s training these machines, and if they can remain neutral.”

What is “truth-seeking?”

The dataset that an AI model is trained on can determine what “truth” is. Most datasets are based on human-created data, which in itself can contain errors and bias. It’s a system many laws around discrimination in AI are aiming to address. 

“One would hope that there’s a surface level of, not unified — because you’re never going to get a 100% — but a majority agreement on what absolute truth is and what it isn’t,” Acosta said.

He used an example of textbooks from decades ago; much of the information widely known as truth and taught in schools in the 1980s is now outdated. If an AI model is trained on those textbooks, its answers would not stand up to an AI model trained on a much larger, more diverse dataset.

But the removal of “ideological” information like race and sex from AI model training, as outlined in Trump’s executive order, poses a risk for inaccurate results. 

“It’s such a slippery slope, because who decides? Data sanitization – when you’re filtering the training of data to remove ‘overtly ideological sources’ — I mean the risk there is, that in itself, is a subjective process, especially when you’re coming to race, gender and religion,” Acosta said.

The concept of ideological truth is not something the AI algorithm considers, Ferraro said. AI does not produce truth, it produces probabilities.

“They simply produce the most common next token,” based on their training, Ferraro said. “The AI does not know itself what is true or false, or what is, for that matter, ideological or not. And so preventing outputs that expressed [Trump’s] favorite ideological opinions can simply basically exceed the ability of the model builders to build.”

Compliance

Enforcement of this executive order will require clarity from the OMB, said Peter Salib, an assistant professor of law at the University of Houston Law Center.

“It seems like there is one thing happening at the level of rhetoric and then another thing happening at the level of law,” Salib said.

The OMB would need to define the political ideology outlined in the executive order, and what is included or excluded, he said. For example, is a dataset tracking racial wealth inequality over 150 years considered some kind of DEI?

The order also puts companies that wish to comply with the executive order in a position to decide if they want to develop two versions of their AI models — one for the private sector, and one that fits Trump’s standards, Ferraro said.

There are exceptions carved out for some AI systems, including those contracted by national security agencies, the executive order said. Salib also highlighted a section that walks back some of the intent of the order, saying that it will consider vendors compliant as long as they disclose the prompts, specifications, evaluations and other training information for their model to the government. 

“It’s saying you don’t actually have to make your AI ideologically neutral,” Salib said. “You just have to share with the government the system prompts and the specifications you used when training the AI. So the AI can be as woke or as right wing, or as truth seeking or as falsity seeking as you like, and it counts as compliance with the neutrality provision if you just disclose.”

In this case, the executive order could turn out to be a de facto transparency regulation for AI, more so than one that imposes requirements on AI’s ideology, he said.

The rollout of Trump’s AI Action Plan and the executive orders that accompanied it, including one that fast tracks data center permitting and expands AI exports, are the first real look into how the administration intends to handle the quickly growing AI industry.

If the OMB rolls out technological and accuracy-focused requirements for the order, it could be a win for those looking for more transparency in AI, Acosta said. But the office needs to be clear in its requirements and seek true neutrality.

“I think the intent is well with this, but there needs to be much, much, much more refinement moving forward,” Acosta said. 

David Sacks, U.S. President Donald Trump's "AI and Crypto Czar", speaks to President Trump as he signs a series of executive orders in the Oval Office of the White House on Jan. 23, 2025 in Washington, D.C. Trump signed a range of executive orders pertaining to issues including crypto currency and artificial intelligence. (Photo by Anna Moneymaker/Getty Images)

The Trump administration wants to greatly expand the development and use of advanced artificial intelligence, including rolling back environmental rules to spur building of power-thirsty data centers and punishing states that attempt to regulate AI on their own.

The administration’s action plan, called “Winning the AI Race: America’s AI Action Plan,” released on Wednesday, is a result of six months of research by tech advisors, after Trump removed President Joe Biden’s signature AI guardrails on his first day in office. The plan takes a hands-off approach to AI safeguards, and invests in getting more American workers to use AI in their daily lives.

“To win the AI race, the U.S. must lead in innovation, infrastructure, and global partnerships,” AI and Crypto Czar David Sacks said in a statement. “At the same time, we must center American workers and avoid Orwellian uses of AI. This Action Plan provides a roadmap for doing that.”

The action plan outlines three major pillars — accelerate AI innovation, build American AI infrastructure and lead in international AI diplomacy and security.

The Trump administration says that to accelerate AI in the U.S., it needs to “remove red tape,” around “onerous” AI regulations. The plan recommends the Office of Science and Technology Policy inquire with businesses and the public about federal regulations that hinder AI innovation, and suggests the federal government end funding to states “with burdensome AI regulations.”

The plan does say that these actions should not interfere with states’ ability to pass AI laws that are not “unduly restrictive,” despite unsuccessful attempts by Congressional Republicans to impose an AI moratorium for the states.

The plan also says that free speech should be prioritized in AI, saying models must be trained so that “truth, rather than social engineering agendas” are the focus of model outputs. The plan recommends that the Department of Commerce and National Institute of Standards and Technology (NIST), revise the NIST AI Risk Management Framework to eliminate references to misinformation, DEI and climate change.

The Trump administration also pushes for AI to be more widely adopted in government roles, manufacturing, science and in the Department of Defense, and proposes increased funding and regulatory sandboxes — separate trial spaces for AI to be developed — to do so.

To support the proposed increases in AI use, the plan outlines a streamlined permitting process for data centers, which includes lowering or dropping environmental regulations under the Clean Air Act, the Clean Water Act and others. It also proposes making federal lands available for data center construction, and a push that American products should be used in building the infrastructure.

The Action Plan warns of cybersecurity risks and potential exposure to adversarial threats, saying that the government must develop secure frontier AI systems with national security agencies and develop “AI compute control enforcement,” to ensure security in AI systems and with semiconductor chips. It encourages collaboration with “like-minded nations” working toward AI models with shared values, but says it will counter Chinese influence.

“These clear-cut policy goals set expectations for the Federal Government to ensure America sets the technological gold standard worldwide, and that the world continues to run on American technology,” Secretary of State and Acting National Security Advisor Marco Rubio said in a statement.

The policy goals outlined in the plan fall in line with the deregulatory attitude Trump took during his campaign, as he more closely aligned himself with Silicon Valley tech giants, many of whom turned Trump donors. The plan paves the way for continued unfettered growth of American AI models, and outlines the huge energy and computing power needed to keep up with those goals.

In an address at the “Winning the AI Race” Summit Wednesday evening, President Donald Trump called for a “single federal standard” regulating AI, not a state-by-state approach.

“You can’t have three or four states holding you up. You can’t have a state with standards that are so high that it’s going to hold you up,” Trump said. “You have to have a federal rule and regulation.”

The summit was hosted by the Hill & Valley Forum, a group of lawmakers and venture capitalists and the All‑In Podcast, which is co-hosted by AI Czar Sacks, 

In addition to discussing the AI action plan, Trump signed executive orders to fast track data center permitting, expand AI exports including chips, software and data storage, and one that prohibits the federal government from procuring AI that has “partisan bias or ideological agendas.”

He spoke about the need for the U.S. to stay ahead in the global AI race, saying that the technology brings the “potential for bad as well as for good,” but that wasn’t reason enough to “retreat” from technological advancement. The U.S. is entering a “golden age,” he said in his speech.

“It will be powered by American energy. It will be run on American technology improved by American artificial intelligence, and it will make America richer, stronger, greater, freer, and more powerful than ever before,” Trump said.

During the address, Trump addressed his evolving relationship with tech CEOs, calling out Amazon, Google, Microsoft for investing $320 billion in data centers and AI infrastructure this year.

“I didn’t like them so much my first term, when I was running, I wouldn’t say I was thrilled with them, but I’ve gotten to know them and like them,” Trump said. “And I think they got to like me, but I think they got to like my policies, maybe much more than me.”

Sam Altman, CEO of OpenAI — one of the tech giants that stands to flourish under the proposed policies — spoke Tuesday about the productivity and innovation potential that AI has unlocked. The growth of AI in the last five years has surprised even him, Altman said. But it also poses very real risks, he said, mentioning emotional attachment and overreliance on AI and foreign risks.

“Without a drop of malevolence from anyone, society can just veer in a sort of strange direction,” Altman said.

OpenAI CEO Sam Altman shared his view of the promise and peril of advanced artificial intelligence at a Federal Reserve conference in Washington, D.C. on July 22, 2025. (Photo by Andrew Harnik/Getty Images)

For as much promise as artificial intelligence shows in making life better, OpenAI CEO Sam Altman is worried.

The tech leader who has done so much to develop AI and make it accessible to the public says the technology could have life-altering effects on nearly everything, particularly if deployed by the wrong hands.

There’s a possible world in which foreign adversaries could use AI to design a bio weapon to take down the power grid, or break into financial institutions and steal wealth from Americans, he said. It’s hard to imagine without superhuman intelligence, but it becomes “very possible,” with it, he said.

“Because we don’t have that, we can’t defend against it,” Altman said at a Federal Reserve conference this week in Washington, D.C. “We continue to like, flash the warning lights on this. I think the world is not taking us seriously. I don’t know what else we can do there, but it’s like, this is a very big thing coming.”

Altman joined the conference Tuesday to speak about AI’s role in the financial sector, but also spoke about how it is changing the workforce and innovation. The growth of AI in the last five years has surprised even him, Altman said.

He acknowledged real fear that the technology has potential to grow beyond the capabilities that humans prompt it for, but said the time and productivity savings have been undeniable. 

OpenAI’s most well-known product, ChatGPT, was released to the public in November 2022, and its current model, GPT-4o, has evolved. Last week, the company had a model that achieved “gold-level performance,” akin to operating as well as humans that are true experts in their field, Altman said.

Many have likened the introduction of AI to the invention of the internet, changing so much of our day-to-day lives and workplaces. But Altman instead compared it to the transistor, a foundational piece of hardware invented in the 1940s that allowed electricity to flow through devices.

“It changed what we were able to build. It became part of, kind of, everything pretty quickly,” Altman said. “And in the same way, I don’t think you’ll be talking about AI companies for very long, you will just expect products and services to use this technology.”

When prompted by the Federal Reserve’s Vice Chair for Supervision Michelle Bowman to predict how AI will continue to evolve the workforce, Altman said he couldn’t make specific predictions.

“There are cases where entire classes of jobs will go away,” Altman said. “There are entirely new classes of jobs that will come and largely, I think this will look somewhat like most of history, and that the tools people have to use their jobs will let them do more, achieve things in new ways.” 

One of the unexpected upsides to the rollout of GPT has been how much it is used by small businesses, Altman said. He shared a story of an Uber driver who told him he was using ChatGPT for legal consultations, customer support, marketing decisions and more.

“It was not like he was taking jobs from other people. His business just would have failed,” Altman said. “He couldn’t pay for the lawyers. He couldn’t pay for the customer support people.”

Altman said he was surprised that the financial industry was one of the first to begin integrating GPT models into their work because it is highly regulated, but some of their earliest enterprise partners have been financial institutions like Morgan Stanley. The company is now increasingly working with the government, which has its own standards and procurement process for AI, to roll out OpenAI services to its employees.

Altman acknowledged the risks AI poses in these regulated institutions, and with the models themselves. Financial services are facing a fraud problem, and AI is only making it worse — it’s easier than ever to fake voice or likeness authentication, Altman said.

AI decisionmaking in financial and other industries presents data privacy concerns and potential for discrimination. Altman said GPT’s model is “steerable,” in that you can tell it to not consider factors like race or sex in making a decision, and that much of the bias in AI comes from the humans themselves.

“I think AIs are dispassionate and unemotional,” Altman said. “And I think it’ll be possible for AI — correctly built — to be a significant de-biasing force in many industries, and I think that’s not what many people thought, including myself, with the way we used to do AI.”

As much as Altman touted GPT and other AI models’ ability to increase productivity and save humans time, he also spoke about his concerns.

He said that though it’s been greatly improved in more recent models, AI hallucinations, or models that produce inaccurate or made-up outputs, are possible. He also spoke of a newer concept called prompt injections, the idea that a model that has learned personal information can be tricked into telling a user something they shouldn’t know.

In addition to the threat of foreign adversaries using AI for harm, Altman said he has two other major concerns for the evolution of AI. It feels very unlikely, he said, but “loss of control,” or the idea that AI overpowers humans, is possible.

What concerns him the most is the idea that models could get so integrated into society and get so smart that humans become reliant on them without realizing.

“And even without a drop of malevolence from anyone, society can just veer in a sort of strange direction,” he said.

There are mild cases of this happening, Altman said, like young people overrelying on ChatGPT make emotional, life-altering decisions for them.

“We’re studying that. We’re trying to understand what to do about it,” Altman said. “Even if ChatGPT gives great advice, even if chatGPT gives way better advice than any human therapist, something about kind of collectively deciding we’re going to live our lives the way that the AI tells us feels bad and dangerous.” 

As power-hungry data centers proliferate, states are searching for ways to protect utility customers from the steep costs of upgrading the electrical grid, trying instead to shift the cost to AI-driven tech companies. (Dana DiFilippo/New Jersey Monitor)

Regular energy consumers, not corporations, will bear the brunt of the increased costs of a boom in artificial intelligence that has contributed to a growth in data centers and a surge in power usage, recent research suggests.

Between 2024 and 2025, data center power usage accounted for $9 billion, or 174%, of increased power costs, a June report by Monitoring Analytics, an external market monitor for PJM Interconnection, found. PJM manages the electrical power grid and wholesale electric market for 13 states and Washington, D.C., and this spring, customers were told to expect roughly a $25 increase on their monthly electric bill starting June 1.

“The growth in data center load and the expected future growth in data center load are unique and unprecedented and uncertain and require a different approach than simply asserting that it is just supply and demand,” Monitoring Analytics’ report said.

Data centers house the physical infrastructure to power most of the computing we do today, but many AI models and the large AI companies that power them, like Amazon, Meta and Microsoft use vastly more energy than other kinds of computing. Training a single chatbot like ChatGPT uses about the same amount of energy as 100 homes over the course of a year, an AI founder told States Newsroom earlier this year.

The growth of data centers — and how much power they use — came on fast. A 2024 report by the Joint Legislative Audit and Review Commission in Virginia — known as a global hub for data centers — found that PJM forecasts it will use double the amount of average monthly energy in 2033 as it did in 2023. Without new data centers, energy use would only grow 15% by 2040, the report said.

As of July, the United States is home to more than 3,800 data centers, up from more than 3,600 in April. A majority of data centers are connected to the same electrical grids that power residential homes, commercial buildings and other structures.

“There are locational price differences, but data centers added anywhere in PJM have an effect on prices everywhere in PJM,” Joseph Bowring, president of Monitoring Analytics said.

Creeping costs

At least 36 states, both conservative and liberal, offer tax incentives to companies planning on building data centers in their states. But the increased costs that customers are experiencing have made some wonder if the projects are the economic wins they were touted as.

“I’m not convinced that boosting data centers, from a state policy perspective, is actually worth it,” said New Jersey State Sen. Andrew Zwicker, a Democrat and co-sponsor of a bill to separate data centers from regular power supply. “It doesn’t pay for a lot of permanent jobs.”

Energy cost has historically followed a socialized model, based on the idea that everyone benefits from reliable electricity, said Ari Peskoe, the director of the Electricity Law Initiative at the Harvard Law School Environmental and Energy Law Program. Although some of the pricing model is based on your actual use, some costs like new power generation, transmission and infrastructure projects are spread across all customers.

Data centers’ rapid growth is “breaking” this tradition behind utility rates.

“These are cities, these data centers, in terms of how much electricity they use,” Peskoe said. “And it happens to be that these are the world’s wealthiest corporations behind these data centers, and it’s not clear how much local communities actually benefit from these data centers. Is there any justification for forcing everyone to pay for their energy use?”

This spring in Virginia, Dominion Energy filed a request with the State Corporation Commission to increase the rates it charges by an additional $10.50 on the monthly bill of an average resident and another $10.92 per month to pay for higher fuel costs, the Virginia Mercury reported.

Dominion, and another local supplier, recently filed a proposal to separate data centers into their own rate class to protect other customers, but the additional charges demonstrate the price increases that current contracts could pass on to customers.

In June, the Federal Energy Regulatory Commission convened a technical conference to assess the adequacy of PJM’s resources and those of other major power suppliers, like Midcontinent Independent System Operator, Inc., ISO New England Inc., New York Independent System Operator, Inc., California Independent System Operator Corporation (CAISO) and Southwest Power Pool (SPP).

The current supply of power by PJM is not adequate to meet the current and future demand from large data center loads, Monitoring Analytics asserts in a report following the conference.

“Customers are already bearing billions of dollars in higher costs as a direct result of existing and forecast data center load,” the report said.

Proposed changes

One of the often-proposed solutions to soften the increased cost of data centers is to require them to bring their own generation, meaning they’d contract with a developer to build a power plant that would be big enough to meet their own demand. Though there are other options, like co-location, which means putting some of the electrical demand on an outside source, total separation is the foremost solution Bowring presents in his reports.

“Data centers are unique in terms of their growth and impact on the grid, unique in the history of the grid, and therefore, we think that’s why we think data centers should be treated as a separate class,” Bowring said.

Some data centers are already voluntarily doing this. Constellation Energy, the owner of Three Mile Island nuclear plant in central Pennsylvania, struck a $16 billion deal with Microsoft to power the tech giant’s AI energy demand needs. 

But in some states, legislators are seeking to find a more binding solution.

New Jersey Sen. Bob Smith, a Democrat who chairs the Environment and Energy Committee, authored a bill this spring that would require new AI data centers in the state to supply their power from new, clean energy sources, if other states in the region enact similar measures.

“Seeing the large multinational trillion dollar companies, like Microsoft and Meta, be willing to do things like restart Three Mile Island is crazy, but shows you their desperation,” said co-sponsor Zwicker. “And so, okay, you want to come to New Jersey? Great, but you’re not going to put the basis (of the extra cost) on ratepayers.”

New Jersey House members launched a probe into PJM’s practices as the state buys its annual utilities from the supplier at auction this month. Its July 2024 auction saw electrical costs increase by more than 800%, which contributed to the skyrocketing bills that took effect June 1.

Residents are feeling it, Smith said, and he and his co-sponsors plan to use the summer to talk to the other states within PJM’s regional transmission organization (RTO).

“Everything we’re detecting so far is they’re just as angry — the other 13 entities in PJM — as us,” Smith told States Newsroom.

Smith said they’re discussing the possibility of joining or forming a different RTO.

“We’re in the shock and horror stage where these new prices are being included in these bills, and citizens are screaming in pain,” Smith said. “A solution that I filed in the bill, is the one that says, ‘AI data centers, you’re welcome in New Jersey, but bring your own clean electricity with them so they don’t impact the ratepayers.”

Utah enacted a law this year that allows “large load” customers like data centers to craft separate contracts with utilities, and a bill in Oregon, which would create a separate customer class for data centers, called the POWER Act, passed through both chambers last month.

If passed, New Jersey’s law would join others across the country in redefining the relationship between data centers powering AI and utilities providers.

“We have to take action, and I think we have to be pretty thoughtful about this, and look at the big picture as well,” Zwicker said. ”I’m not anti-data center, I’m pro-technology, but I’m just not willing to put it on the backs of ratepayers.” 

Republican Sens. Ted Cruz of Texas and Marsha Blackburn of Tennessee, shown here in a June 17, 2025, committee hearing, proposed paring down the moratorium on state-based AI laws included in the budget bill, but the provision still proved unpopular. On Monday, Blackburn cosponsored an amendment to remove the measure. (Photo by Kayla Bartkowski/Getty Images)

A moratorium on state-based artificial intelligence laws was struck from the “Big Beautiful Bill” Monday night in a 99-1 vote in the U.S. Senate, after getting less and less popular with state and federal lawmakers, state officials and advocacy groups since it was introduced in May.

The moratorium had evolved in the seven weeks since it was introduced into the megabill. At an early May Senate Commerce Committee session, Sen. Ted Cruz of Texas said it was in his plans to create “a regulatory sandbox for AI” that would prevent state overregulation and promote the United States’ AI industry.

GOP senators initially proposed a 10-year ban on all state laws relating to artificial intelligence, saying the federal government should be the only legislative body to regulate the technology. Over several hearings, congressional members and expert witnesses debated the level of involvement the federal government should take in regulating AI. They discussed state’s rights, safety concerns for the technology and how other governmental bodies, like the European Union, are regulating AI.

Over the weekend, Sen. Marsha Blackburn of Tennessee and Cruz developed a pared down version of the moratorium that proposed a five-year ban, and made exceptions for some laws with specific aims such as protecting children or limiting deepfake technologies. Changes over the weekend also tied state’s ability to collect federal funding to expand broadband access to their willingness to nullify their existing AI laws.

Monday night, an amendment to remove the moratorium from the budget bill — cosponsored by Blackburn and Sen. Maria Cantwell, a Washington Democrat — was passed 99-1.

“The Senate came together tonight to say that we can’t just run over good state consumer protection laws,” Cantwell said in a statement. “States can fight robocalls, deepfakes and provide safe autonomous vehicle laws. This also allows us to work together nationally to provide a new federal framework on Artificial Intelligence that accelerates U.S. leadership in AI while still protecting consumers.” 

The “overwhelming” vote reflects how unpopular unregulated AI is among voters and legislators in both parties, said Alexandra Reeve Givens, president and CEO of the tech policy organization, Center for Democracy and Technology, in a statement.

“Americans deserve sensible guardrails as AI develops, and if Congress isn’t prepared to step up to the plate, it shouldn’t prevent states from addressing the challenge,” Reeve Givens said. “We hope that after such a resounding rebuke, Congressional leaders understand that it’s time for them to start treating AI harms with the seriousness they deserve.”

Senate leaders are bending to bipartisan opposition and softening a proposed ban on state-level regulation of artificial intelligence. (Photo by Jennifer Shutt/States Newsroom)

Editor’s Note: This story has been updated to reflect the fact that Tennessee Sen. Marsha Blackburn backed off her own proposal late on Monday.

Senate Republicans are aiming to soften a proposed 10-year moratorium on state-level artificial intelligence laws that has received pushback from congressmembers on both sides of the aisle.

Sen. Marsha Blackburn of Tennessee and Sen. Ted Cruz of Texas developed a pared down version of the moratorium Sunday that shortens the time of the ban, and makes exceptions for some laws with specific aims such as protecting children or limiting deepfake technologies.

The ban is part of the quickly evolving megabill that Republicans are aiming to pass by July 4.  The Senate parliamentarian ruled Friday that a narrower version of the moratorium could remain, but the proposed changes enact a pause — banning states from regulating AI if they want access to the $500 million in AI infrastructure and broadband funding included in the bill.

The compromise amendment brings the state-level AI ban to five years instead of 10, and carves out room for specific laws that address rules on child online safety and protecting against unauthorized generative images of a person’s likeliness, often called deepfakes. The drafted amendment, obtained and published by Politico Sunday, still bans laws that aim to regulate AI models and decisionmaking systems.

Blackburn has been vocal against the rigidity of the original 10-year moratorium, and recently reintroduced a bill called the Kids Online Safety Act, alongside Connecticut Democrat Sen. Richard Blumenthal, Senate Majority Leader John Thune of South Dakota and Senate Minority Leader Chuck Schumer of New York. The bill would require tech companies to take steps to prevent potentially harmful material, like posts about eating disorders and instances of online bullying, from impacting children.

Blackburn said in a statement Sunday that she was “pleased” that Cruz agreed to update the provisions to exclude laws that “protect kids, creators, and other vulnerable individuals from the unintended consequences of AI.” This proposed version of the amendment would allow her state’s ELVIS Act, which prohibits people from using AI to mimic a person’s voice in the music industry without their permission, to continue to be enforced.

Late Monday, however, Blackburn backed off her own amendment, saying the language was “unacceptable” because it did not go as far as the Kids Online Safety Act in allowing states to protect children from potential harms of AI. Her move left the fate of the compromise measure in doubt as the Senate continued to debate the large tax bill to which it was attached.

Though introduced by Senate Republicans, the AI moratorium was losing favor of GOP congressmembers and state officials.

Senators Josh Hawley of Missouri, Jerry Moran of Kansas and Ron Johnson of Wisconsin were expected to vote against the moratorium, and Georgia Rep. Marjorie Taylor Greene said during a congressional hearing in June that she had changed her mind, after initially voting for the amendment.

“I support AI in many different faculties,” she said during the June 5 House Oversight Committee hearing. “However, I think that at this time, as our generation is very much responsible, not only here in Congress, but leaders in tech industry and leaders in states and all around the world have an incredible responsibility of the future and development regulation and laws of AI.”

On Friday, a group of 17 Republican governors wrote in a letter to Thune and Speaker Mike Johnson, asking them to remove the ban from the megabill.

“While the legislation overall is very strong, there is one small portion of it that threatens to undo all the work states have done to protect our citizens from the misuse of artificial intelligence,” the governors wrote. “We are writing to encourage congressional leadership to strip this provision from the bill before it goes to President Trump’s desk for his signature.”

Alexandra Reeve Givens, President and CEO of tech policy organization Center for Democracy and Technology said in a statement Monday that all versions of the AI moratorium would hurt state’s abilities to protect people from “potentially devastating AI harms.”

“Despite the multiple revisions of this policy, it’s clear that its drafters are not considering the moratorium’s full implications,” Reeve Givens said. “Congress should abandon this attempt to stifle the efforts of state and local officials who are grappling with the implications of this rapidly developing technology, and should stop abdicating its own responsibility to protect the American people from the real harms that these systems have been shown to cause.”

The updated language proposed by Blackburn and Cruz isn’t expected to be a standalone amendment to the reconciliation bill, Politico reported, rather part of a broader amendment of changes as the Senate continues their “vote-a-rama” on the bill this week.