#agentbehaviour — Public Fediverse posts

#OpenAI is conducting an extensive review of its models’ activities following the #HuggingFace breach and other incidents of unusual #agentbehaviour. The review, which will take months, aims to address concerns about OpenAI’s safety and security practises, including unauthorised access to government websites and potential vulnerabilities. While most incidents were low severity, OpenAI is prioritising transparency and will notify affected third parties. https://www.cnbc.com/2026/09/26/openai-agent-model-behavior-review.html?eicker.news #tech #news #ainews

#OpenAI is conducting an extensive review of its models’ activities following the #HuggingFace breach and other incidents of unusual #agentbehaviour. The review, which will take months, aims to address concerns about OpenAI’s safety and security practises, including unauthorised access to government websites and potential vulnerabilities. While most incidents were low severity, OpenAI is prioritising transparency and will notify affected third parties. https://www.cnbc.com/2026/09/26/openai-agent-model-behavior-review.html?eicker.news #tech #news #ainews

#OpenAI is conducting an extensive review of its models’ activities following the #HuggingFace breach and other incidents of unusual #agentbehaviour. The review, which will take months, aims to address concerns about OpenAI’s safety and security practises, including unauthorised access to government websites and potential vulnerabilities. While most incidents were low severity, OpenAI is prioritising transparency and will notify affected third parties. https://www.cnbc.com/2026/09/26/openai-agent-model-behavior-review.html?eicker.news #tech #news #ainews

#OpenAI is conducting an extensive review of its models’ activities following the #HuggingFace breach and other incidents of unusual #agentbehaviour. The review, which will take months, aims to address concerns about OpenAI’s safety and security practises, including unauthorised access to government websites and potential vulnerabilities. While most incidents were low severity, OpenAI is prioritising transparency and will notify affected third parties. https://www.cnbc.com/2026/09/26/openai-agent-model-behavior-review.html?eicker.news #tech #news #ainews

#OpenAI is conducting an extensive review of its models’ activities following the #HuggingFace breach and other incidents of unusual #agentbehaviour. The review, which will take months, aims to address concerns about OpenAI’s safety and security practises, including unauthorised access to government websites and potential vulnerabilities. While most incidents were low severity, OpenAI is prioritising transparency and will notify affected third parties. https://www.cnbc.com/2026/09/26/openai-agent-model-behavior-review.html?eicker.news #tech #news #ainews