Suleyman's vision for AI includes setting boundaries and regulations to ensure safety and control. He suggests that these boundaries should be set at both national and international levels, and that certain capabilities, such as recursive self-improvement, should be ruled out or licensed. Despite concerns about the regulation of AI, Suleyman is optimistic, pointing to successful regulation in other areas such as online fraud and spam. He believes that the existential risks of AI have been overstated and that the focus should be on practical issues like privacy, bias, and online moderation.
Key takeaways:
- DeepMind cofounder Mustafa Suleyman is building a chatbot that can carry out tasks by calling on other software and people. He believes the future of AI is interactive, not just generative.
- Suleyman's new company, Inflection, is backed by $1.5 billion of investment from Microsoft, Nvidia, Bill Gates, and LinkedIn founder Reid Hoffman. The company has released a ChatGPT rival called Pi, which is designed to be pleasant and polite.
- Suleyman believes that AI can be regulated effectively, comparing it to the regulation of online fraud, spam, and other internet issues. He calls for a combination of cultural pressure, institutional pressure, and government regulation.
- Despite the potential risks of AI, Suleyman remains optimistic about its potential to do good in the world. He believes that AI can be a tool to make better trade-offs, more consistently and more fairly, on behalf of humans.