TOOL LETS YOU CHECK IF YOU'RE IN LLM TRAINING DATA

As more internet traffic flows into large language models, questions mount about what personal data ends up in AI training datasets. A new project addresses this concern head-on by letting users search their own recognition across multiple models. The site, built by a design partnership over recent weeks, works by querying frontier and small-scale LLMs simultaneously. It clusters their responses and provides metrics on recognition strength—essentially measuring how much of your digital presence was absorbed during training. The tool targets a growing gap in AI transparency. With training datasets increasingly drawn from public internet sources, many users lack visibility into whether their content contributed to model development. The parallel querying approach offers broader coverage than single-model searches. The project gained traction on Hacker News with 103 points and 76 comments, indicating strong interest in AI data provenance issues. Creators indicate openness to user questions about methodology and results.