Helen Toner calls for federal disclosure rules, third-party audits and “emergency intervention powers” over frontier AI
Executive director of Georgetown’s Center for Security and Emerging Technology and a former OpenAI board member, interviewed on Offline with Jon Favreau
“I think what we should be doing now is setting ourselves up to be able to navigate that better. So what does that mean? It means… anything around more transparency from the company, so disclosure requirements… federal requirements around sharing what kind of AIs they’re training, what kind of tests they’re running, why they think what they’re doing is safe, what their security practices are… I do think the third-party auditors is an important element… Would love to see better kind of incident reporting but also investigations… you could have kind of a … government entity that is able to actually go in … demand information. I think that is where I would start.” … “I actually think more useful than a kill switch is kind of emergency intervention powers of different kinds… giving… government the ability to intervene in the case of an emergency… could be helpful. I think having clearer standards on what is okay development practices here.” … “We also need to have kind of these regulatory frameworks… looking internally at their development practices as well.”
Source and context
Original video
About this source
Weekly podcast, published on YouTube at 9:00 a.m. EDT on September 19, 2026. Host Jon Favreau interviews Helen Toner, executive director of Georgetown’s CSET and a former OpenAI board member.
Archived copy (opens in a new tab)Before the quotation
Host Jon Favreau asked what Toner thought would make the most difference if Congress passed AI legislation over the next year or two, noting he did not need her to be too specific.
After the quotation
Favreau then asked whether limits could be placed on AI systems’ recursive self-improvement. Toner said it was “worth exploring” but that she did not yet know how to write it into a statute.
How this statement is classified
Case context: Should Washington impose stronger safeguards on frontier AI?
The label describes this statement’s response within the context above.
Why this label?
The case’s fixed reference point is whether Washington should impose stronger safeguards on frontier AI. Toner proposes several specific new federal requirements — disclosure rules, third-party audits, incident-investigation powers, emergency-intervention authority and oversight of internal development practices — which supports the proposition. The closest competing label is mixed_or_conditional, because she is skeptical of a literal “kill switch” as one specific proposal; it loses because she reframes rather than rejects that idea, proposing broader emergency-intervention powers in its place, and every other element of her answer adds to rather than qualifies the case for stronger safeguards.
- Recorded on
- Published here
Should Washington impose stronger safeguards on frontier AI?
Explore the case context, sources and public responses.
More from this case
Read the full case“I will not hinder or stifle the growth of AI.”Read statement
“I joined @RepYassAnsari and more than two dozen @HouseDemocrats in urging Donald Trump to take urgent action to address the serious risks posed by frontier AI.”Read statement
“This idea that if we regulate AI, that China will win, is not true. And the way that we know that is because China regulates AI. They have the most regulation of their AI than any country in the world... So the tech bros in the U.S. that say if we regulate AI, the Chinese going to win - that's a self-serving argument. That isn't true.”Read statement