Documents show OpenAI lobbied for parts of the EU's AI Act to be watered down, including successfully avoiding its general purpose AI being deemed “high risk”
The CEO of OpenAI, Sam Altman, has spent the last month touring world capitals where, at talks to sold-out crowds …
The documents add evidence that OpenAI’s regulatory engagement was not only about advocating balanced rules in public forums, but also about shaping the scope of obligations applied to general-purpose AI.
First-order effects
OpenAI avoids having its general-purpose AI treated as “high risk” under the proposed framework, reducing its immediate exposure to the stricter treatment associated with that designation.
EU policymakers and stakeholders must assess the Act with a clearer record of how a major model provider sought changes to its provisions.
Second-order effects
Other frontier-model providers gain a concrete incentive to press for risk rules that distinguish general-purpose systems from higher-risk deployments.
The episode raises the importance of lobbying capacity and direct policymaker access in determining which AI firms can help shape compliance boundaries.
Third-order effects
If this pattern persists, AI regulation may increasingly be set through negotiation between governments and a small group of well-resourced labs, rather than through rules applied uniformly after products reach market.
The lasting policy question is whether governance can separate legitimate technical input from regulatory outcomes that entrench incumbent model providers.
The trend: This is one data point in the rise of the state-compatible AI lab, where frontier developers treat regulatory engagement as a core competitive function.
🚨 SCOOP: OpenAI lobbied the E.U. to weaken forthcoming AI regulation, even as in public it calls for stronger AI guardrails, documents obtained by TIME show. We're publishing the key document in full alongside my story 👇 https://time.com/...
@profnoahgian
Noah Giansiracusa on x
I suspect this is going to be one of the biggest tech stories of 2023—it confirms what many of us have been suspecting all along and shows how the sausage is really made. https://twitter.com/...
@marleenstikker
@marleenstikker on x
This is exactly the reason why the AI ‘existential risk for humanity’ tech bros are fearmongering our societies. To keep us busy with demystifying nonsens and bogus science while they are lobbying intensively against regulation. https://time.com/...
@schock
@schock on x
You can't trust these companies any further than you can throw a server rack full of GPUs. “OpenAI lobbied the E.U. to weaken forthcoming AI regulation, even as in public it calls for stronger AI guardrails, documents obtained by TIME show” https://time.com/...
@colincrowell
Colin Crowell on x
Meet the new Big Tech boss...same as the old Big Tech boss 🇪🇺🤖🤷⤵️ https://twitter.com/...
@garymarcus
Gary Marcus on x
Powerful reminder that we simply cannot trust Big Tech. A must read. https://twitter.com/...
@djleufer
Daniel Leufer on x
Me & @sarahchander spoke to Billy about this very interesting document from OpenAI In autumn 2022 a disturbing loophole appeared in the Council text of the AI Act, exempting “purely accessory” AI systems from all requirements It's very in line with what OpenAI demanded https://tw…
@abebab
Abeba Birhane on x
people like @sama ask to be regulated in public. behind doors, they lobby for the opposite. watch what these folks do not what they say https://twitter.com/...
@npcollapse
Connor Leahy on x
No way, who could have expected this? Great reporting Billy! https://twitter.com/...
@josephjacks_
@josephjacks_ on x
The regulatory capture industrial complex (led by Microsoft / OpenAI) is in FULL swing ... https://twitter.com/...
@tprstly
Theo on x
Sam Altman is not your friend. https://time.com/...