The prospect of the government vetting AI models is causing a minor panic in certain tech policy and industry circles.
This week, the
New York Times and
POLITICO reported that the White House is considering an executive order that would, in part, require government testing and approval before AI companies can release their models.
Kevin Hassett
then said during a Wednesday Fox Business appearance that the executive order may include “a clear roadmap to everybody about … how future AIs that also potentially create vulnerabilities should go through a process so that they’re released to the wild after they’ve been proven safe, just like an FDA drug.”
The news came as the administration met with tech companies about Mythos and other AI models that potentially pose a dire
threat to cybersecurity. Such a proposal for the government to pre-approve model releases has received
opposition from free-market think tanks and industry-aligned tech policy advisers, along with some academics.
Much of the debate speaks to a question that any policymaker faces when trying to regulate a potentially harmful product: Do you try to ensure something is safe before it’s released, or punish companies after the fact if something goes wrong?
When it comes to tech regulation, the U.S. has traditionally favored the latter approach of imposing penalties after a product causes harm in the wild.
“It gets to what is the heart and soul of America’s technological governance regime, which is that we have a regime for computing and digital communication that’s much more ex-post and responsive in character,” said Adam Thierer, a resident senior fellow at the R Street Institute.
There are plenty of reasons to think that going to the other extreme of requiring government pre-approval would be a bad idea for AI.
“Frankly, it would create a false sense of security,” said Neil Chilson, a former acting chief technologist at the Federal Trade Commission who is currently the Abundance Institute’s head of AI policy. “There’s only so much you can do in a lab environment … to actually study something that is a general-purpose technology, that is going to be deployed in lots of different spaces.”
Getting upfront government approval could be an issue if it leads people to be less vigilant about problems after release. “You also want to be continuing to iterate and modify and monitor after that release,” said Encode General Counsel Nathan Calvin.
Given the notoriously glacial pace of bureaucracy, waiting on the government to give the green light could slow innovation, but also make it harder for public watchdogs to conduct their own oversight.
“A lot of the most concerning risks from frontier models come from how they are used internally within the companies that develop them,” Helen Toner, interim executive director at the Georgetown Center for Security and Emerging Technology, told DFD. “Creating extra barriers to releasing them would do nothing to solve that problem, and would lengthen the time period when companies have advanced models internally that the public — and independent experts — have no access to.”
And if you consider models to have some elements of speech, then there could be constitutional issues with a pre-approval regime. The First Amendment forbids “prior restraint,” which involves requiring government authorization before something can be expressed. American laws around defamation or obscenity are designed to punish illicit speech after the fact. Prohibiting prior restraint is a core tenet of First Amendment law, born out of a history of censors attempting to control
printing presses. Thierer suggests that having the government oversee model releases could run afoul of this principle.
“Information communication technologies have special First Amendment concerns … when you regulate them,” he said. “If you’re regulating a model to be ‘safe,’ that can be done along certain dimensions where you could probably try to avoid getting into speech concerns, but at some point you might.”
Yet if models like Mythos are as powerful as companies claim, then waiting for harm to occur could result in a catastrophe that’s hard to mitigate. Even so, Toner suggests that looping in regulators on an advisory level prior to release could be a preferable alternative to giving them veto power. “[I]t obviously would have been better for CISA to have early access to Mythos and be actively working to coordinate cyber defense among U.S. critical infrastructure providers,” she said.
This is more aligned with the approach of the Center for AI Standards and Innovation at the Commerce Department. The agency announced deals with Microsoft, xAI and Google DeepMind
on Tuesday to conduct pre-deployment evaluations, though this is done on a voluntary basis. Chilson was supportive of the idea: “[Companies] don’t always have all of the context for what the government might know about our adversaries or what they might know about our own governmental systems.”
A White House official told DFD on background: “Any policy announcement will come directly from the President. Discussion about potential executive orders is speculation.”
Coalition of civil society, tech policy, and voter engagement groups launch MA voter petition campaign urging Rep. Trahan to stop preemption
On Thursday, a coalition of civil society, tech, and voter engagement organizations launched a new voter petition campaign urging Rep. Lori Trahan (MA-03) not to cosponsor legislation banning state AI laws. The petition effort, which has already gathered over 3,800 signatures from Massachusetts voters, is supported by Americans for Responsible Innovation (ARI), Demand Progress, Design It For Us, Encode, Indivisible (Acton Area, Massachusetts), and Public Citizen. The latest effort comes on the heels of statements from Massachusetts state lawmakers who have urged the Massachusetts Democrat not to cosponsor preemption legislation.
“Massachusetts state lawmakers have spent years working on safeguards that protect children, families, and workers in the AI era. Freezing those state laws would be a huge mistake, handing Big Tech immunity from accountability,” said ARI VP of Communications Chris MacKenzie. “Thousands of voters in Massachusetts are speaking up because they don’t want another era of ‘move fast and break things.’ We should be looking for ways to put commonsense guardrails in place, not writing industry a blank check.”
“Big Tech CEOs and the White House keep trying to ban states from enacting AI safeguards and Rep. Trahan must not give them what they want,” said Demand Progress AI Policy Advisor Colin McGlynn. “State laws are one of the few defenses we have left against unaccountable AI chatbots that have sexual conversations with children, encourage people in crisis to harm themselves and even give instructions on how to carry out a mass shooting. Rep. Trahan must not put Big Tech profits over the safety of Americans by supporting AI preemption legislation.”
“As a young person based in Massachusetts, I know firsthand the anxieties young people in the state are feeling with unregulated artificial intelligence,” said Kaitlynn Cherry, Policy Lead, Design It For Us. “We cannot allow the federal government to greenlight AI amnesty, especially when states across the country are stepping up to act. We urge Rep. Trahan to heed the calls of young people from Massachusetts and reject any AI legislation that wipes out states’ rights to regulate in this area.”
“There is overwhelming support for stronger guardrails around the development and use of AI,” said Adam Billen, Co-Executive Director, Encode AI. “Massachusetts has been a national leader in addressing AI-related risks. Freezing those efforts now would roll back the hard-won protections Massachusetts lawmakers have established and leave the state’s children, families, and workers more vulnerable to harm from AI now and into the future. We urge Rep. Trahan to put the safety of her constituents ahead of corporate interests and oppose this preemption legislation.”
“Our Congresswoman is negotiating with House Republicans for a federal bill that threatens to wipe out our state’s AI defenses and hand immunity to Big Tech,” said Stella Ko of Indivisible, Acton Area. “If this passes, we lose our ability to protect residents from deepfakes, data breaches, and election interference. This is a line in the sand. We must speak up now to stop it.”
“Big Tech corporations want to end state AI regulation because they know states like Massachusetts have led the way in advancing common-sense safeguards and restraints on Big Tech abuses,” said Public Citizen Co-President Rob Weissman. “They would love nothing more than modest federal rules – which will quickly grow stale and likely not be updated – to stamp out the ability of states to respond quickly to emerging harms from AI and Big Tech. Rep. Trahan cares about protecting kids and consumers – which is exactly why she should immediately abandon any consideration of federal preemption of state AI rules.”
###
Americans for Responsible Innovation (ARI) is a nonprofit organization dedicated to policy advocacy in the public interest, focused on emerging technologies like artificial intelligence (AI). Learn more at ARI.us.