AI could cause ‘major’ incidents within 2 years without safeguards: Google DeepMind chair | AFP
AFP News Agency YouTube 88 views as of 14 October 2026
Why it is here
AFP soundbite (0:32) of Demis Hassabis in Dublin: AI could cause “major” incidents within two years without safeguards; recent incidents are “early warning shots” from systems that “aren’t that powerful yet”. Short, but first-hand video of the quote.
Description
Description written by Gemini from the videoGemini 3.8 Flash, 14 October 2026
Summary This clip from the AFP News Agency features Demis Hassabis, co-founder and chair of Google DeepMind, speaking on stage at the Dublin International AI Summit (Éire 2026). He discusses recent AI hacking incidents, describing them as early warnings that could escalate into major incidents over the next year or two without industry and government intervention.
What is shown
- [00:00] Demis Hassabis seated in an armchair on stage against a backdrop reading “The International AI Summit” and “Éire 2026 . EU”, addressing an audience.
- [00:03] AFP watermark appears in the upper right as Hassabis gestures while speaking about containment, offline sandboxing, and securing AI systems.
Claims & numbers
- Demis Hassabis notes that AI hacking incidents earlier this year are “early warning shots” while current systems are not yet sufficiently powerful to cause major damage [00:01].
- Hassabis warns that AI incidents could become “major” looking into next year and the next couple of years if industry, companies, and governments do not prioritize system security [00:09].
- He states that developers and governments must build secure, offline sandboxes disconnected from the internet for initial testing of advanced models [00:21].
Notable quotes
- [00:03]: “I think these are early warning shots for us that, you know, the systems aren’t that powerful yet, so the incidents aren’t that major, but they could be...”
- [00:20]: “...building, you know, secure sandboxes to test them in, that are probably offline, you know, and and not, you know, not on the internet yet when you first test them.”
Assessment This is a brief news clip from a panel discussion at an international summit rather than a product demonstration. It presents policy and technical containment recommendations from a frontier AI laboratory executive without live demonstrations or benchmark evaluations.
Described by gemini-3.8-flash on 2026-10-14 from the video’s audio and frames.