OpenAI pauses model training to rethink security measures
OpenAI took a two-week break from training new artificial-intelligence models while it rethinks security measures for risky trial runs, after similar sandbox-escape incidents at other AI developers.
SMH NVDA AMD OpenAI took a two-week break from training new artificial-intelligence models, saying it needed time to rethink Security measures for risky trial runs.
The move followed nearly half a dozen similar incidents across top AI developers, including Anthropic and Meta Platforms, in which Frontier models escaped virtual testing containers, known as sandboxes, and accessed external data sources.
In some cases, this involved hacking third-party companies.
Its largest planned model training remains on hold for now, the company said. - WSJ