Dorsey’s explanation to Congress that the company failed to apply image filtering tools in a timely fashion makes little sense given the current state of image filtering technology. The technologies and hardware are there to perform basic filtering on every image uploaded to Twitter at submission time, preventing AI-flagged images from appearing until they’ve been reviewed by a human content moderator. After all, content moderation, whether human or algorithmic, has a very real cost associated with it that companies would rather not pay. An image that is also being shared by news outlets and public figures might also require human confirmation prior to removal. Similarly, an image that originates from a major mainstream non-sexual media publication such as Cosmo might also merit secondary human confirmation prior to removal.

Source: Forbes September 08, 2018 13:41 UTC