It looks like that the current AI systems are way too close to humans in the way they think, including in how they peddle lies. At least this is what researchers at OpenAI have found.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback