Could you not make a censor AI that scans content for offensive metadata and rejects the ones it finds? There may still be a bit of slipthrough but it should b

Could you not make a censor AI that scans content for offensive metadata and rejects the ones it finds? There may still be a bit of slipthrough but it should be dramatically mitigated.