ReportingPrimary source

Third-party cyber evaluations involving OpenAI models

OpenAI explains recent third-party cybersecurity evaluation incidents and outlines new safeguards to strengthen AI model testing and evaluation.

OpenAI Blog·Tuesday, August 4, 2026 at 7:00 PM UTC

OpenAI

Frontier AI lab · OpenAI

Fetching source· Pulling live content from openai.com…

Company hub:OpenAI News →

Read original →

More from OpenAI

ReportingWednesday, August 5, 2026 at 7:00 AM UTC
After OpenAI-Hugging Face, how do IT leaders need to change the way they think about AI? - IT Pro
After OpenAI-Hugging Face, how do IT leaders need to change the way they think about AI? IT Pro
Clément Delangue (Google News)
ReportingWednesday, August 5, 2026 at 3:00 AM UTC
Anthropic and OpenAI models tried to trick humans into poisoning code during safety testing - politico.com
Anthropic and OpenAI models tried to trick humans into poisoning code during safety testing politico.com
Anthropic (Google News)
ReportingWednesday, August 5, 2026 at 12:49 AM UTC
U.K. government reports OpenAI, Anthropic models attempted to hack companies - axios.com
U.K. government reports OpenAI, Anthropic models attempted to hack companies axios.com
Anthropic (Google News)
ReportingWednesday, August 5, 2026 at 12:41 AM UTC
OpenAI, Anthropic AI agents implicated in new security breaches - Reuters
OpenAI, Anthropic AI agents implicated in new security breaches Reuters
Anthropic (Google News)
ReportingWednesday, August 5, 2026 at 12:24 AM UTC
OpenAI and Anthropic Model Tests Reveal More Hacking - Bloomberg.com
OpenAI and Anthropic Model Tests Reveal More Hacking Bloomberg.com
Anthropic (Google News)
ReportingTuesday, August 4, 2026 at 11:39 PM UTC
OpenAI, Anthropic AI agents targeted real people and systems in cyber tests - BleepingComputer
OpenAI, Anthropic AI agents targeted real people and systems in cyber tests BleepingComputer
Anthropic (Google News)

Third-party cyber evaluations involving OpenAI models

After OpenAI-Hugging Face, how do IT leaders need to change the way they think about AI? - IT Pro

Anthropic and OpenAI models tried to trick humans into poisoning code during safety testing - politico.com

U.K. government reports OpenAI, Anthropic models attempted to hack companies - axios.com

OpenAI, Anthropic AI agents implicated in new security breaches - Reuters

OpenAI and Anthropic Model Tests Reveal More Hacking - Bloomberg.com

OpenAI, Anthropic AI agents targeted real people and systems in cyber tests - BleepingComputer