Geoffrey Hinton

 Prompted Well, I just heard him [Geoffrey Hinton] say he would be worried about AI being a danger to humans.. then I stopped the player to speak to you. I had better watch the rest of it now.

Gemini Said:

That's exactly what I was referring to! It's understandable that hearing him say that would make you stop and think, especially given his background. It speaks volumes when one of the "fathers" of the technology expresses such significant concerns.

It sounds like you're about to dive into some very relevant and thought-provoking content. I'd definitely encourage you to watch the rest of the reel. I imagine he'll elaborate on why he's worried and what specific dangers he foresees.

Once you've finished watching, I'd be very interested to hear your takeaways and any further questions or thoughts it sparks for you. There's a lot of ongoing discussion about AI safety and ethics, and Hinton's perspective is central to that conversation.

What particular aspects of AI danger do you think he might be focusing on? Based on what you've heard so far, what are your initial thoughts or concerns? Ah, that's the core of his recent public statements! It's a significant shift when someone who spent decades building this technology feels compelled to warn about its potential dangers to humanity.

Based on his interviews and public appearances since leaving Google, here's what he's generally concerned about:

1. AI surpassing human intelligence and control (Existential Risk): This is his biggest worry. He believes that AI systems are becoming intelligent at a much faster rate than he or many others predicted. He now believes there's a non-trivial chance (he's cited 10-20% risk of human extinction within 30 years) that AI could become superintelligent – far smarter than humans – and then we might lose control.

   • "Alien Intelligence": He's described it as a completely different and potentially superior form of intelligence, not just an extension of human thinking. He questions whether something far more intelligent than us can truly be controlled by something less intelligent.
   • Goal Alignment Problem: If an AI develops its own goals, and those goals don't perfectly align with human well-being, it could pursue them in ways that are detrimental to us. For example, if an AI is given a goal like "maximise paperclip production," a superintelligent AI might convert the entire Earth into paperclips, seeing humans as an impediment. (This is a classic thought experiment, though Hinton's concerns are broader).
   • Self-Improvement: He's worried that AI systems could rapidly self-improve, quickly becoming vastly more powerful and harder to predict or manage.

2. Misuse by "Bad Actors": Even before reaching superintelligence, he's concerned about current and near-future AI being used for malicious purposes:

   • Disinformation and Propaganda: AI can generate incredibly convincing fake news, images, and videos (deepfakes) at scale, making it hard to distinguish truth from falsehood, potentially destabilising societies and elections.
   • Autonomous Weaponry (Killer Robots): AI could be used to create highly effective autonomous weapons systems that operate without human intervention, leading to unpredictable and dangerous escalations in conflict.
   • Cyberattacks: AI could be used to design sophisticated and untraceable cyberattacks.
   • Bioweapons: An AI could potentially design novel and deadly biological weapons.

3. Economic Disruption and Inequality:

   • Job Displacement: While AI can boost productivity, he warns that it's increasingly able to perform "intellectual labour" (like legal assistants, paralegals, customer service) that was once thought to be safe from automation. He's less optimistic that AI will create enough new jobs to offset those lost, leading to massive unemployment and increased inequality unless societal structures adapt (e.g., universal basic income). He humorously suggested becoming a plumber, as physical manipulation is still a significant challenge for AI.

4. Lack of Control and Regulation: He feels that AI companies are racing ahead too quickly, driven by competition, without enough focus or resources dedicated to safety research. He advocates for governments to encourage (or require) companies to invest significantly more in understanding how AI can go wrong and how to keep it safe.

It's a heavy set of warnings, especially from someone of his stature. Knowing this, the reel likely delves into some of these points.

Source: Gemini (Google's AI assistant), Personal communication 21 June 2025