SEDUCTIVE AI: People prefer the empathy of humans, but rate ‘fake’ AI empathy higher.

UPDATE: From the comments:

This is going to be a problem. Already is starting to be. Truly skilled first-rate human flatterer/persuader/seducers are rare, and generally end up aimed at high payoff targets. Most people in their whole lives will never have rated more than passing attention from such, and thus will have little or no immunity.

And now that level of individually customized flatterer/persuader/seducer skills is being automated, and will soon be cheaply deployable against *everyone*. I foresee big trouble.

I address these very issues in my forthcoming Seductive AI.