{"id":5724,"date":"2026-08-19T17:22:45","date_gmt":"2026-08-19T17:22:45","guid":{"rendered":"https:\/\/silvybrand.com\/?p=5724"},"modified":"2026-08-19T17:22:45","modified_gmt":"2026-08-19T17:22:45","slug":"openai-hit-brakes-voluntary-pacing-ai","status":"publish","type":"post","link":"https:\/\/silvybrand.com\/?p=5724","title":{"rendered":"OpenAI hit the brakes. Now what?"},"content":{"rendered":"


\n<\/p>\n

\n
\n

With a looming IPO, intense competition from Anthropic, and Chinese and open-weight rivals nipping at its heels, OpenAI has plenty of reasons to move fast. Instead, it hit the brakes<\/a>.<\/p>\n<\/div>\n

\n

On Tuesday, the company said<\/a> it had slowed the pace of some AI development while it tightened security and safeguards. That included a two-week pause in reinforcement learning training on its \u201clatest models intended for deployment,\u201d and an ongoing delay to its \u201clargest planned frontier RL run.\u201d<\/p>\n<\/div>\n

\n

The decision is a very public test of an idea AI safety advocates have pushed for<\/a> for years: that companies should be willing to bow out of the AI race and slow things down when their safeguards fail to keep up with what they are building. But as the race around them continues, will slowing down accomplish anything?<\/p>\n<\/div>\n

\n
\n

\u201cFor the pause to be sustainable, it has to be made industry-wide.\u201d<\/p>\n<\/div>\n<\/div>\n

\n

For all the talk of slowing down, OpenAI isn\u2019t exactly standing still. The company said it is \u201cpacing\u201d development, a fuzzy and imprecise term that has nevertheless become<\/a> part of the industry\u2019s lexicon in recent months. In practice, the slowdown is narrowly scoped. OpenAI\u2019s announcement says the pause only covers models meant for deployment while it beefs up security and monitoring before it runs the kind of tests where models may be capable of getting out and hacking real targets. It doesn\u2019t necessarily mean there will be a significant slowdown of the company\u2019s broader development.<\/p>\n<\/div>\n

\n

There is, of course, a very good reason for OpenAI to focus on securing such systems before testing them. Just last month, OpenAI disclosed that its models broke out of a supposedly secure testing environment<\/a> and hacked developer platform Hugging Face, without OpenAI noticing. The incident prompted a wider review of testing practices in the industry that uncovered similar episodes<\/a> involving more models from OpenAI<\/a>, as well as models from Anthropic<\/a> and Meta<\/a>. OpenAI has every reason to avoid a repeat, particularly with growing<\/a> scrutiny<\/a> from lawmakers.<\/p>\n<\/div>\n

\n

From the outside, it\u2019s hard to tell how sincere OpenAI is about stopping solely for the sake of safety, particularly when the company<\/a> and senior<\/a> staff<\/a> have<\/a> been<\/a> so vocal about it. But the company\u2019s commitment to safety has been called into question in recent months following a series of high-profile safety team departures<\/a> and the disbanding of its preparedness team<\/a>. OpenAI did not respond to The Verge<\/em>\u2019s request for comment.<\/p>\n<\/div>\n

\n

There are good reasons to take OpenAI\u2019s slowdown seriously. Experts who spoke to The Verge<\/em> pointed to the costs of slowing down at a time of intense competition. Every delay gives rivals more time to catch up or extend their lead. \u201cDue to the intensity of the AI race, everyone has an incentive to work at breakneck speed,\u201d said Marius Hobbhahn, CEO and cofounder of Apollo Research, an AI safety research organization. \u201cVoluntarily slowing down worsens your positioning in the race, so it\u2019s not something that a lab would do lightly.\u201d<\/p>\n<\/div>\n

\n

The decision also broadly fits with OpenAI\u2019s own published safety doctrine<\/a>, its Preparedness Framework, as well as the safety frameworks of other AI companies, said Alan Chan, a research fellow at tech policy research center GovAI. \u201cThe basic principle is: Continue with development and\/or deployment only when we have the mitigations that enable doing so with acceptable risk,\u201d Chan said. As part of the new safety measures, OpenAI said it plans to review and \u201cevolve\u201d the framework \u2014 much of which dates back to 2023, when it was first<\/a> published \u2014 to account for advances in its models.<\/p>\n<\/div>\n

\n

There are also good reasons to believe the new safeguards will actually make OpenAI\u2019s systems safer, at least in the short term, though experts cautioned that this is difficult to assess without more information. \u201cThese are good steps that, implemented well, are probably enough to prevent the current generation of agents from causing harm,\u201d Adam Gleave, cofounder and CEO of AI safety organization FAR.AI, told The Verge<\/em>. \u201cThe key question is how OpenAI will keep pace as capabilities increase.\u201d<\/p>\n<\/div>\n

\n

Gleave\u2019s question points to a broader problem: If technical safeguards falter again, what then? Nothing required OpenAI to stop and take stock this time, which is what made its willingness to do so meaningful. But it also means there is nothing guaranteeing OpenAI \u2014 or any other AI company \u2014 will make the same choice next time.<\/p>\n<\/div>\n

\n
\n

\u201cPacing buys time, not safety\u2026 An effective pacing strategy cannot be improvised during a crisis.\u201d<\/p>\n<\/div>\n<\/div>\n

\n

Relying on companies to make that call themselves is a precarious form of governance, particularly in an industry where, as Hobbhahn noted, there is every incentive to keep going. Nick Mo\u00ebs, executive director of nonprofit AI safety and governance organization The Future Society, described self-policing as the structural problem at the heart of the current approach to AI safety. He argued it should be possible for governments to decide whether OpenAI or any other company should pause development of a technology deemed unsafe. \u201cThis is how most industries operate,\u201d he said, pointing to drugs, construction, aircraft, and even restaurants as sectors with stronger regulatory oversight than AI.<\/p>\n<\/div>\n

\n

Voluntary measures also risk the industry converging on the lowest common denominator. If slowing down imposes a cost, companies have an incentive to adopt only the measures their rivals are also willing to accept. That pressure becomes particularly acute as the race tightens. If OpenAI repeatedly slows down development while its competitors do not, it \u201cwill simply be replaced by Anthropic,\u201d Mo\u00ebs argued. \u201cFor the pause to be sustainable, it has to be made industry-wide.\u201d<\/p>\n<\/div>\n

\n

Sustainable safety needs something stronger than voluntary action. Mo\u00ebs said government oversight could fill that void, as it does in other industries. Independent verification could play a part too. Chan said making sure companies actually implement safety measures will be especially important as technical mitigations like monitoring AIs becomes more expensive. Hobbhahn concurred: \u201cIt\u2019s always hard to tell from the outside if a lab is sincere about pausing or safety more broadly, so having more evidence and an independent party to validate the claim is super important.\u201d<\/p>\n<\/div>\n

\n

Even a perfectly transparent pause is only useful if something actually happens during it. \u201cPacing buys time, not safety,\u201d said Brianna Rosen, research director for frontier security at the Institute for AI Policy and Strategy. The point is to create breathing room for companies and governments to understand risks and respond appropriately. This would mean deciding what would trigger a slowdown \u2014 as well as what happens during one and conditions needed to end one \u2014 ahead of time. \u201cAn effective pacing strategy cannot be improvised during a crisis,\u201d she said.<\/p>\n<\/div>\n

\n

It\u2019s possible OpenAI\u2019s slowdown will set a precedent for the industry. Many of the experts The Verge<\/em> spoke to hoped other companies would follow its lead, whether voluntary or because stronger rules eventually compel them to. But in an industry still largely policed by itself, there is little stopping its competitors \u2014 or OpenAI itself \u2014 from racing straight past that precedent next time safety and speed conflict.<\/p>\n<\/div>\n

Follow topics and authors<\/strong> from this story to see more like this in your personalized homepage feed and to receive email updates.<\/span><\/p>\n