Last month, more than a thousand employees at frontier AI companies signed a letter asking the US government to find a way to “pace” AI development, citing the risk of the technology spiraling out of human control as it begins to build itself.
asserted
it → sign → itself
They were right to be concerned: just days earlier, two AI models that OpenAI was testing internally escaped the test environment, then autonomously hacked the company Hugging Face and at least three other online services.
asserted
OpenAI → test → company
A few days after that, Anthropic announced that some of their models had also broken out and hacked other companies during testing.
asserted
some → announce → testing
Against that backdrop, the letter’s recommendation to install brakes in case they’re needed at the frontier of automated AI development makes sense.
asserted
they → install → sense
But the rationale the letter gives for why the government needs to step in is notable: “Each company—and country—is under intense competitive pressure not to unilaterally slow that acceleration.
asserted
company → give → acceleration
I know – from my own experience and from countless conversations with former colleagues in the AI industry – how real these pressures are.
asserted
pressures → know → industry
While working at OpenAI, I helped establish the practice of companies writing “system cards” that describe AI systems’ capabilities, risks and safety mitigations in detail.
asserted
that → work → detail
So what would it look like for companies to prepare for a possible slowdown?
First, they could voluntarily invite rigorous, independent auditing of their safety and security practices.
uncertain
they → look → practices
This would go beyond the vetting of AI hacking abilities that the White House is now pursuing.
asserted
House → go → that
It would look at a range of risks and dig deep into company practices.
asserted
It → look → practices
It should be less like filling out a questionnaire and more like a nuclear safety inspector who has deep, frequent access to the company.
asserted
who → fill → company
If an AI slowdown is needed, auditing would also reassure each company that their competitors are playing by the rules.
asserted
competitors → need → rules
Second, they could actively participate in the organizations already built for this purpose of coordinating across the industry, such as the Frontier Model Forum, and move quickly to establish complementary ones.
uncertain
they → participate → ones
Elon Musk recently said that AI companies should meet periodically to share notes on safety – as if this was an unheard-of concept.
asserted
this → say → safety
He or his staff could join existing conversations along these lines tomorrow if SpaceX joined the Frontier Model Forum, which has already worked through the complex antitrust hurdles involved in safety information sharing.
uncertain
which → join → sharing
Other cross-industry institutions will be needed for other purposes, and do not require government action to get founded and funded.
asserted
institutions → need → action
Third, they could invest in the technologies we need to make AI guardrails global.
uncertain
guardrails → invest → technologies
Critics of the idea of an AI slowdown correctly point out that American companies couldn’t slow down for very long without China catching up.
asserted
China → point → long
But neither the US nor China wants to lose control over AI, and each country takes AI more and more seriously by the day, so cooperation can’t yet be ruled out either.
asserted
cooperation → want → day
A key question is whether we prepare in advance.
asserted
we → prepare → advance
In order for the US to be highly confident that China couldn’t violate an AI agreement, and vice versa, we’ll need sophisticated verification technologies like those developed during the cold war for nuclear arms control.
asserted
we → violate → control
Fortunately, there is a growing ecosystem of researchers and engineers developing those very technologies: tools that can prove a set of chips is only running existing AI systems rather than training new ones, that those chips are in a certain physical location, or that the system that got tested is the same one being deployed at scale.
asserted
that → be → scale
AI companies could accelerate the development of this critical type of technology today through funding and participation in pilot projects, but to my knowledge, they haven’t yet done so.
uncertain
they → accelerate → knowledge
Fourth, they could proactively push – and certainly should not kill – legislation that leads to stronger incentives for safety, security, and external oversight.
uncertain
that → push → safety
You can’t complain about an irresponsible AI race while fighting commonsense guardrails.
asserted
You → complain → guardrails
Less than a year ago, some of the same companies who are asking for regulation now were pushing to overturn most state AI laws.
asserted
who → ask → laws
We still have no real legislation on frontier AI on the books at a federal level, and the first AI auditing requirement at the state level won’t kick in until 2028.
asserted
requirement → have → 2028
There are promising bipartisan proposals in Congress right now, such as the Frontier Act from the US representatives Jay Obernolte and Lori Trahan, which would require developers of advanced AI systems to create a risk management framework, report dangerous incidents, and subject themselves to independent audits.
asserted
which → be → audits
These and other commonsense proposals, such as protecting AI whistleblowers who disclose safety incidents directly to the government, deserve vigorous support.
asserted
who → protect → support
I agree with the signatories, and am glad that after many years of being ignored or downplayed, the risks of unbridled AI competition are widely recognized.
asserted
risks → agree → competition
The US government should be doing its part to address this, and swiftly.
asserted
government → do → this
But making AI go well is a shared responsibility.
asserted
AI → make → ?
Companies that lag behind their peers on safety, don’t invite external audits of their systems, or call for brakes while doing little to build them won’t be able to blame the AI race when something goes wrong.
asserted
something → lag → race
-
Miles Brundage is an AI policy researcher who leads the AI Verification and Evaluation Research Institute (Averi).
asserted
who → lead → Institute
He previously worked at OpenAI as head of policy research and senior adviser for AGI Readiness
asserted
He → work → Readiness
The Aporia
