Futurology

I recall hearing about this - the AI did this by actively lying to users, inventing experiences, e.g. pretending to be a male victim of SA or a poc opposed to black lives matter, etc.

I’m not surprised, but still doing this study was unethical. Even good research, if done unethically, is bad science.

Turns out, when you’re willing to lie about nearly everything about yourself, you can convince people of things. I’m absolutely shocked. Completely taken aback. No one could have possibly predicted that con men could be automated.