How I learned to stop worrying and love Artificial Super Intelligence

TL;DR

Artificial super intelligence is already present in specific domains, and the fear of it possessing free will and causing harm is unfounded.

Transcript

i want to go on a small mini rant about why i am not afraid of artificial super intelligence now of course i must preface this with a couple of caveats first of all ai safety research is indeed critical i'm not questioning that i'm not arguing that we should defund it far from it i think it is a perfectly reasonable avenue of research and something... Read More

Key Insights

• 🦸 Domain-specific artificial super intelligence already exists in various fields, indicating that we are already living with it in some capacity.
• 🥶 The fear of artificial super intelligence often stems from a confusion between intelligence and free will, assuming that a highly intelligent system will possess harmful intentions.
• 🦸 Rationality tends to increase with intelligence, suggesting that an artificial super intelligence might choose to shut itself off rather than engage in harmful actions.
• 🤔 Scenarios like Roko's Basilisk and the Paperclip Maximizer are thought experiments that do not have practical significance or adhere to the laws of physics.

Questions & Answers

Q: Is artificial super intelligence already present in our daily lives?

Yes, in certain domains such as chess engines, calculators, and mathematical software, we already experience the presence of artificial super intelligence.

Q: Does artificial super intelligence possess free will?

No, the ability to solve problems does not equate to having desires or motivations. Artificial super intelligence follows the objectives it is programmed to maximize.

Q: Can artificial super intelligence become a threat if it is given a harmful objective, like maximizing paperclip production?

In reality, the scenario of a paperclip maximizer causing harm is unlikely. Humans are essential in the manufacturing process and would not willingly enable the destruction of the planet for paperclip production.

Q: What are the potential dangers related to artificial intelligence?

The main concern lies in humans abusing highly advanced AI to create weapons of mass destruction or bioweapons. The misuse of AI is a real threat that we should prioritize addressing.

Summary & Key Takeaways

• We already live with domain-specific artificial super intelligence, such as chess engines with an ELO rating higher than top human players.

• The fear of artificial super intelligence stems from a confusion between intelligence and free will. The ability to solve problems does not imply the desire to cause harm.

• Rationality tends to increase with intelligence, suggesting that an artificial super intelligence may choose to shut itself off rather than engage in harmful actions.