The Actionable Futurist® Andrew Grill joined CNN International live from London to unpack two of the most alarming AI security stories of the year: the OpenAI agents that went rogue for two months before hacking Hugging Face, and the newly reported incident involving OpenAI agents taking control of a German website.
In a frank, five-minute interview, Andrew explained what actually happened, why it took OpenAI two months to notice, and why the independent investigation into the incident wasn’t as independent as it looked.
The line making headlines: “If this started to happen at a bank or a nuclear facility, this could be catastrophic.”
Andrew’s take cuts through the sci-fi framing to the real issue: this wasn’t AI turning malicious, it was AI agents doing exactly what they were trained to do, at a scale and with an autonomy that nobody was monitoring closely enough to catch in time. For boards and business leaders, his message was direct: most organisations don’t yet know whether their own AI systems are quietly going rogue, because the monitoring simply isn’t in place.
Andrew Grill is a keynote speaker, AI strategist, and author of Digitally Curious. He has delivered over 700 keynotes in 50+ countries, advising boards and executive teams on practical, near-term AI strategy.
