Demis Hassabis urges caution: AI should not prioritize speed and disruption

Demis Hassabis, the co-founder of Google DeepMind, a major AI company in the UK, expressed his belief that the AI industry should avoid the approach taken by older tech giants. He referred to Mark Zuckerberg's motto, "Move fast and break things," which originally aimed to promote rapid innovation and company growth but later came to symbolize reckless behavior from big tech firms. Hassabis argued that this mindset should not be applied to AI due to its critical importance. He emphasized the need for understanding and responsible deployment of AI systems, calling for extensive work in this area. Speaking ahead of the UK's AI safety summit, Hassabis outlined three broad categories of threats posed by artificial intelligence that require varying solutions and immediate attention. DeepMind is currently developing AlphaFold, an AI program that has the potential to revolutionize the discovery of new medicines by accurately predicting protein structures in the human body. The company achieved a significant milestone in 2016 when its previous product, AlphaGo, defeated the world's top human player in the strategic game Go, leading to the player's retirement. Tech leaders are advocating for government regulations to govern the rapidly advancing technology. The AI safety summit, hosted at Bletchley Park, where World War Two codebreakers operated, has brought together around 100 world leaders, tech executives, academics, and AI researchers. The objective of the summit is to discuss strategies for maximizing the benefits of AI, such as in medicine and climate change solutions, while minimizing risks. The event will primarily focus on the extreme threats posed by cutting-edge AI, with an emphasis on bio-terrorism and cyber-attacks. Notable attendees include US Vice President Kamala Harris, European Commission President Ursula von der Leyen, and representatives from China. Critics have raised concerns about the dominance of US tech giants in the guest list, including OpenAI, Anthropic, Microsoft, Google, Amazon, Tesla, and X (previously Twitter). Prime Minister Rishi Sunak will also engage in a livestream conversation with Elon Musk on the topic of AI. Despite some criticism and overshadowing by recent AI safety announcements from the US and G7, Hassabis believes the UK can still play a significant role in shaping the discussions. Aidan Gomez, the founder of Cohere, expressed his opinion that there are more immediate threats in AI that should be addressed, rather than focusing on the "doomsday Terminator scenario." He emphasized the need for concrete policy work and regulation to ensure the responsible use of AI.