AI researchers put out videos saying superintelligence is ‘exactly as dangerous as it sounds’
The Verge Stevie Bonifield ● Covered by 2 sources
AI researchers are posting interviews warning superintelligence could kill us. One former OpenAI and DeepMind staffer says human extinction looks like a coin flip.
Based on reporting by The Verge, Stevie Bonifield — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
A group of AI researchers has started putting their worst-case fears on video, and they are not being subtle about it. In interviews collected by Palisade Research and published on frominside.ai, several current and former employees from OpenAI, Google, Anthropic and Google DeepMind warned that advanced AI could wipe out humanity.
The bluntest line came from Geoffrey Irving, who used to work at OpenAI and Google DeepMind. He said the chance of human extinction is “about a coin flip,” in his view. That is not the language of cautious lab notes. It is a public warning, and it lands hard because it comes from someone who has worked inside the companies building these systems.
Neel Nanda, a research scientist at Google DeepMind, was less absolute but still grim. He said there is “at least a 10 percent chance” that AI leads to human extinction, according to the source article. That kind of number is still enormous when the thing being measured is whether people survive their own invention.
Palisade Research describes itself as a non-profit studying AI capabilities and motivations. Its interview series is meant to surface what researchers actually think when the PR gloss is gone. And judging by these clips, a growing slice of the field is no longer talking about AI as a productivity tool with rough edges. They are talking about a system they believe could become exactly as dangerous as it sounds.
My take — AI-written commentary, not fact-checked reporting
This is what happens when a field spends years calling everything “transformative” and then acts shocked when some of the people inside it start saying “catastrophic” out loud. The useful part here is not the drama; it is the reminder that the AI industry still likes to sell confidence while its own researchers keep reaching for fire alarms. No amount of glossy demos makes that look normal.
Read more about this at: The Verge