In a unified effort to combat the spread of child sexual abuse material (CSAM), leading technology companies including Google, Meta, OpenAI, Microsoft, and Amazon have pledged to rigorously review and clean their AI training datasets of any CSAM. This commitment is part of a broader initiative established through newly agreed principles aiming to minimize the risks of CSAM proliferation.
These measures include ensuring training data is free of CSAM, avoiding high-risk datasets, and diligently removing any CSAM links or imagery from future model training sources. Additional efforts involve thorough testing of AI models to prevent the generation of CSAM and proceeding with model releases only after thorough child safety evaluations.
This initiative also encompasses contributions from Anthropic, Civitai, Metaphysic, Mistral AI, and Stability AI amid rising concerns around AI-generated fake CSAM content and its implications. The involvement is a response to findings by Stanford researchers and increased pressure on NCMEC due to a surge in AI-generated CSAM reporting. Google has further supported this cause by enhancing ad grants for NCMEC, aiming to elevate public awareness on preventing and reporting abuse.
Source: The Verge
Comments