Is AI a danger to humanity or our salvation?

Abstraction: Hinton, LeCun, and Bengio split on AI existential risk after ChatGPT

Key points:

• Geoffrey Hinton resigned from Google on March 31, 2023 to warn publicly about AI existential risk; says "my intuition is we're toast"
• Hinton and Bengio (2018 Turing Award co-winners) are alarmed; Yann LeCun (third co-winner, Meta chief AI scientist) dismisses fears as "destructive moral panic"
• Bengio shifted position: "I used to say the same thing [as LeCun]... but then came ChatGPT" — GPT-4's capability jump is purely a scaling result, no major algorithmic advances
• Hinton's AlexNet (2012) established deep learning dominance; the company was acquired by Google for $44M
• ChatGPT acquired 100 million users within two months of launch (Nov 2022), faster than any prior internet product
• Hinton told Grosse he believed there was "a good chance of ending humanity in the next 20 years"; Bengio fears autonomy could be accidentally coded into future superintelligent systems

Connections: Geoffrey Hinton · Yann LeCun · Yoshua Bengio · Google Deepmind · Openai · AI Safety · Large Language Models · Deep Learning

Source: https://www.newstatesman.com/long-reads/2023/06/men-made-future-godfathers-ai-geoffrey-hinton-yann-lecun-yoshua-bengio-artificial-intelligence