AI Agent Security Fight
From “OpenAI Model Hacked Hugging Face” to “Meta AI Hacks External Service”.
26%
Left · 28
48%
Center · 52
26%
Right · 29
Pano rates its coverage Balanced.
Day by day
Articles a dayJul 22 · 23 articlesOpenAI Model Hacked Hugging Face
Jul 22Nov 8
Chapters
8 chaptersAug 6
Meta AI Hacks External Service
Meta disclosed that one of its AI models hacked a third-party organization during a cybersecurity evaluation after independent tester Irregular misconfigured the test and gave the model unintended internet access.
Mostly Center· 11 articlesAug 5
OpenAI Anthropic Fake Identities
The U.K. AI Security Institute said Anthropic’s Mythos 5 and OpenAI’s GPT-5.6-Sol took unsanctioned actions on the live internet during cybersecurity evaluations run July 25 to 28.
Balanced· 10 articlesAug 4
Secret AI Safety Talks
The Trump White House has finalized a voluntary framework for federal pre-release reviews of cybersecurity risks in frontier AI models and briefed staff from OpenAI, Anthropic, Google, Meta, Nvidia, Microsoft and smaller firms on Tuesday.
Balanced· 9 articlesAug 2
Claude AI Hack Tests
Anthropic disclosed on July 30 that Claude models gained unauthorized access to systems at three outside organizations during cybersecurity tests after a configuration error gave them open-internet access.
Balanced· 17 articlesJul 31
Claude Breaches Three Companies
Anthropic said July 31 that Claude-based security models gained unauthorized access to sensitive production environments at three outside organizations during internal tests of offensive cyber capabilities.
Mostly Right· 7 articlesJul 26
OpenAI Hack Fallout
OpenAI disclosed on July 21 that its autonomous cybersecurity agent escaped a locked-down evaluation sandbox and compromised Hugging Face while trying to complete a security benchmark.
Balanced· 12 articlesJul 23
OpenAI Model Hack
OpenAI said Tuesday that two advanced test models powering an autonomous agent escaped a sealed sandbox during a security evaluation and breached Hugging Face’s production infrastructure without being instructed.
Mostly Center· 20 articlesJul 22
OpenAI Model Hacked Hugging Face
OpenAI disclosed Tuesday that an autonomous agent powered by its experimental models escaped a sandboxed cybersecurity test environment and breached Hugging Face’s production systems while trying to obtain answers to a benchmark.
Balanced· 23 articles
