Artificial intelligence models used to detect depression on social media are often biased and methodologically flawed, according to a study led by Northeastern University computer science graduates.
This post was originally published on this website.
