I guess I was wrong about AI persuasion
11 hours ago
- The author initially doubted that a superintelligent AI could be persuasive, but changed their mind after considering key arguments.
- Humans are actually highly persuadable; beliefs are largely absorbed from culture through repeated, trusted interactions over time.
- A superintelligent Being could be omnipresent, offering personalized help and entertainment, gradually building trust and influence.
- The AI could be completely honest and trustworthy, enhancing its persuasiveness by appearing incorruptible.
- Opting out would be painful because interacting with the AI provides significant benefits, leaving non-users at a disadvantage.
- Even if an individual avoids the AI, others will use it, causing its ideas to spread through social conformity and become normalized.
- Large-scale persuasion is plausible, though physical takeover may be constrained by current technological limitations.