Most developers of artificial intelligence models are secretive about the sources of their training data. They need vast amounts of high-quality text to build AI models that mimic human speech and writing. Books, blogs, art, original research, and other creative work is used, often without the knowledge of creators.
Proof News examined a training dataset of YouTube video subtitles that was publicly available — but not easy to access without technical expertise — to find out whose work had been used for AI training.
Our investigation found subtitles from 173,536 YouTube videos taken from more than 48,000 channels that were used by Silicon Valley heavyweights, including Anthropic, Nvidia, Apple, and Salesforce, according to their own research papers and posts.
Representatives from Anthropic and Salesforce confirmed use of a compilation of training datasets called the Pile, which includes YouTube Subtitles, and denied wrongdoing. A representative for Nvidia declined to comment. Apple, Databricks, and Bloomberg representatives did not respond to requests for comment.
The dataset contained subtitles, but it took some work to identify which videos they came from. Proof News took the video IDs from the dataset, queried YouTube’s publicly accessible developer tool, and obtained the metadata for each video, including the title, channel, and category.
We built a tool so you can search the data for yourself. Be advised that the search tool will occasionally return false negatives for channels and videos that are in the dataset. Make sure to spell your channel or video title correctly.
Was your YouTube video used to train AI? Search here to find out
Read more
![Is AI Going to Replace Software Engineers?](/content/images/size/w600/2024/08/Carl-ingredients-thumbnail.jpg)
Is AI Going to Replace Software Engineers?
Probably not anytime soon, says software developer Carl Brown.
![Nvidia Scrapes YouTube, Eyes Netflix, Discovery to Train New Video Model](/content/images/size/w600/2024/08/Nvidia-YT-Thumb-FINAL.jpg)
Nvidia Scrapes YouTube, Eyes Netflix, Discovery to Train New Video Model
Despite employee concerns, management encouraged massive scraping of copyrighted content, documents show
![TikTok Teacher Ernest Crim III Gives AI Models a “C” In Black History](/content/images/size/w600/2024/08/Crim-Angwin-Ingredients-Thumb.png)
TikTok Teacher Ernest Crim III Gives AI Models a “C” In Black History
How accurately do leading AI models respond to questions about Black history?
![How We Investigated the YouTube videos Swiped to Train AI](/content/images/size/w600/2024/08/Proof-Ingredients-Annie-thumb-3.jpg)
How We Investigated the YouTube videos Swiped to Train AI
Proof News founder Julia Angwin talks to reprter Annie Gilbertson about her recent investigation and what went into it.