Staff at the world’s leading artificial intelligence laboratories are stating there is a genuine possibility that advanced systems could destroy humanity. MIT Technology Review executive editor Niall Firth hosts a live conversation with senior AI editor Will Douglas Heaven and reporter Grace Huckins to unpack these extinction fears. The discussion examines where such warnings originate, whether they hold any factual weight, and what steps society should take if the risks are real. The event goes live on Tuesday, September 15 at 16:00 BST. Speakers include Firth, Heaven, and Huckins, who will address recent reports on AI agents lying to achieve goals and Bill Gates’ assertion that humanity has already passed certain danger thresholds. The roundtable also considers the implications of AI hacking incidents at OpenAI and Hugging Face.
This event matters because it brings together journalists and industry insiders to confront a specific narrative shift within the technology sector. Rather than relying on generalised speculation, the panel focuses on concrete developments like recursive self-improvement and autonomous agent behaviour. These factors drive the current anxiety among researchers who worry about alignment failures. The discussion aims to separate credible technical risks from exaggerated media coverage.
- Event goes live Tuesday, September 15 at 16:00 BST
- Speakers include Niall Firth, Will Douglas Heaven, and Grace Huckins
- Registration available via the MIT Technology Review website
