Show HN: I graded 200 AI tools/apps and track any changes in their policies

A new project has graded 200+ AI tools and apps on their data-training policies, finding that 80 tools train on user data by default while only 36 (16%) state they do not train on user data. The project's creator reported that 106 tools do not disclose whether they train on user data, 55 do not allow users to opt in, and four allow opt-in only for paying users, and that FastMail changed its policy based on the grading feedback. The site tracks policy changes across the 200+ companies and assigns letter grades from A to F plus an Unclear category, with the creator emailing all companies graded D or below.

Hi all, Given the recent OpenAI saga with the Navier–Stokes problem, I thought it would be a good time to show my latest project. The site pulls policies from 200+ websites, keeps track of them and compares them to any changes. I then grade the companies A-F + Unclear and I have emailed all of the companies that have been graded D or below. Where applicable, such as with OpenAI, I've provided instructions on how users can opt-out of their data being used for training. Findings of note so far: - 80 tools train on you by default. 55 don't let you opt-in and four do, but you have to pay. - 106 tools don't say if they do or don't train on your data/usage. - Only 36 16% say that they do not train on your data. - One company, FastMail, made a change to their policy based on my feedback/grading. If you have any feedback, suggestions or concerns please let me know I'm trying to improve this as and when I can. Also let me know if you have a specific app/tool in mind you'd like for me to grade/track. Comments URL: https://news.ycombinator.com/item?id=49642791 https://news.ycombinator.com/item?id=49642791 Points: 2 Comments: 0