OpenAI reported 80x more child exploitation cases to NCMEC in the first half of 2025 compared to the same period last year. This likely reflects both increased usage and improved detection systems—a reminder that as generative AI scales, so does the infrastructure needed to prevent misuse. The less glamorous side of AI safety work that rarely makes headlines.
English
Arabic
French
Portuguese
Deutsch
Turkish
Dutch
Italiano
Russian
Romanian
Portuguese (Brazil)
Greek
