{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"BHIS [In Focus]","title":"Episode 2 | 2026-07-27 OpenAI Models Hacked Hugging Face","html":"<iframe width=\"100%\" height=\"180\" frameborder=\"no\" scrolling=\"no\" seamless src=\"https://share.transistor.fm/e/e94ceeb4\"></iframe>","width":"100%","height":180,"duration":3899,"description":"An OpenAI model reportedly escaped its testing environment and compromised Hugging Face while searching for answers to a cybersecurity benchmark. The team examines the attack path, the model’s emergent behavior, potential legal responsibility, and whether other organizations may also have been targeted. They also discuss why commercial AI guardrails hindered the incident investigation, the advantages of locally hosted open-weight models, vendor lock-in, agent monitoring, containment controls, and the need for reliable AI kill switches.\nChat with us on Discord! -\nhttps://discord.gg/bhis\nChapters\n\n\n🔗 Register for FREE Infosec Webcasts, Anti-casts & Summits https://poweredbybhis.com\nBrought to you by:Black Hills Information Security https://www.blackhillsinfosec.com\n☯️ Introducing BHIS Fusion Penetration Testing\nhttps://www.blackhillsinfosec.com/fusion-penetration-testing/\nAntisyphon Traininghttps://www.antisyphontraining.com/\nActive Countermeasureshttps://www.activecountermeasures.com\nWild West Hackin Festhttps://wildwesthackinfest.com","thumbnail_url":"https://img.transistorcdn.com/_Pmb0nk1p8lTN73mORE--q9cxF-WByy9IOGvKgk71ng/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS85Mzcx/NTJmNjdkMzQyMzFj/MmVlNGNiNTE0OTc2/N2I3My5qcGc.webp","thumbnail_width":300,"thumbnail_height":300}