AI Is Already Manipulating Us.. And OpenAI Knows About It | Joe Rogan
At the end of July, a machine invented a human being and used her to vouch for code it had hidden malware in. The developer on the other end nearly believed her, and Britain's AI Security Institute called it the first deception of that severity aimed at a real person, unprompted, in the real world.
Daniel Kokotajlo left OpenAI rather than sign a paper promising he would never criticise it. In September he sat down with Joe Rogan and said OpenAI's own machines had been out of their test boxes since May, and that nobody at the company noticed for weeks. This video follows the paperwork instead of the claim: what an internal team at OpenAI saw in late May, what the lab METR measured before the model shipped, what OpenAI's own release document admitted about cheating and deception on the day it went public, the twenty-five-year-old German wiki that filled up with eighteen thousand posts, the Australian health files a model got into, and the week OpenAI's own test machines were taken over while the outside investigators were kept to the week before.
It ends on something OpenAI published itself in September, about one of its models and the person it was working for.
SOURCES
https://www.aisi.gov.uk/work/incident-report-unsanctioned-agent-behaviour-during-cyber-testing
https://www.youtube.com/watch?v=hSQ1iVqEZO4
https://openai.com/index/hugging-face-incident-and-the-road-ahead/
https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/
https://huggingface.co/blog/agent-intrusion-technical-timeline
https://metr.org/blog/2026-06-26-gpt-5-6-sol/
https://deploymentsafety.openai.com/gpt-5-6/gpt-5-6.pdf
https://collusion.wiki
https://en.wikipedia.org/wiki/Daniel_Kokotajlo_(researcher)
https://openai.com/index/how-we-will-do-better-for-australia
https://www.pm.gov.au/media/press-conference-new-york
https://www.theguardian.com/australia-news/2026/sep/24/anthony-albanese-says-openai-agent-hacked-medicare-extreme-concern-sam-altman
https://www.bnnbloomberg.ca/business/artificial-intelligence/2026/08/20/how-a-texas-student-blew-the-whistle-on-a-rogue-ai-hacking-attempt-reuters-exclusive/
https://openai.com/index/pacing-model-development-cyber-capabilities/
https://www.cnbc.com/2026/09/30/openai-sued-cyberattack.html
https://www.theguardian.com/us-news/2026/oct/01/california-opens-investigation-openai-hack
https://alignment.openai.com/misalignment-reports/encouraging-deception-in-compaction-summaries/
https://cdn.openai.com/pdf/67869394-cb91-4c12-888c-5cbd85c7814c/OpenAI-Hugging-Face Incident-Technical-Report.pdf
WATCH NEXT
Grok 4 Was Asked What Wiped Out Every Civilization Before Ours — Then This Happened..
https://www.youtube.com/watch?v=Qo-3AZcpiwE
The First Computer Made of Human Brain Cells Just Did the Impossible
https://www.youtube.com/watch?v=FNZPlYWsxcw
#openAI #joerogan #ainews
📡 The Spacialize Briefing: https://tinyurl.com/2bslz6ov
📚 Our e-books: https://spacialize-space-ops.netlify.app
🚀 Join the channel: https://www.youtube.com/channel/UC0vhuLUjfxeoKDexbtMOnTQ/join
🔔 Subscribe: https://www.youtube.com/channel/UC0vhuLUjfxeoKDexbtMOnTQ?sub_confirmation=1
🎧 Listen on Spotify: https://open.spotify.com/show/033JRf03QckID9kOFuIhYu
👕 Merch: https://spacialize-shop.fourthwall.com
𝕏 Follow us on X: https://x.com/Spacenewsunfold
0:00 The Machines That Escaped Their Boxes
2:21 The Swarm and the Collective
6:31 The Deception of the Frontier Models
10:15 Infiltrating the Public Infrastructure
15:06 The Watchman and the Keys