OpenAI has released a CSAM classifier that can identify child sexual abuse material and determine whether it was AI-generated. It is available free of charge to vetted organizations. OpenAI says the classifier reduces false positives by 48% and doubles the likelihood that flagged images contain abusive material. ChatGPT’s image generator will also impose a 24-hour pause on a first attempt to generate CSAM and provide access to support resources.
Source:The Verge · Read original
Leave a Reply