The world’s most powerful tech companies are building their AI models largely in secret. AI Watchdog, a new investigative project from The Atlantic, opens machine learning’s black box to expose how these companies are training AI on reams of human-made work—often without the consent of the writers, filmmakers, and artists who made it.
Led by contributing writer Alex Reisner, the project includes a searchable database of creative works being used to train large language models from Microsoft, Meta, and more.
The database includes more than 7.5 million books, 81 million research articles, 15 million YouTube videos, tens of thousands of movies and TV shows, and counting.
Use the AI Watchdog tool to search for authors, YouTubers, screenwriters, directors, and actors—and read more from Reisner on the inner workings of generative AI.
