This is specifically supposed to be a "persuasion-detection" system, not generalized phishing detection? Like, it would trigger on a personal email from your friend asking you to support their GoFundMe? Seems like it would be weird to warn someone about any type of persuasion

Super underrated, this is very likely going to happen

@firstuserhere it just doesn't sound that useful... What would you do if you detect persuasion? Warn the user about it? They're reading the email, can't they already see that it's persuasion?

@ErickBall Yep! warn the user about it - that's a lot of utility. No, it's not easy to tell when someone is stating something or pushing an opinion subtly

@ErickBall I'd find sources for it later but OpenAI and google both have had this in works at some stage in 2023

How on earth would you detect that?

Got idea for market from a talk by Ilya Sutskevar (founder OpenAI)