Just realized my AI writing tool was feeding my drafts into a public dataset

I found a line in the terms of service yesterday that said any text I generate gets stored for training purposes. Looked up the dataset on Hugging Face and sure enough, there were snippets of my emails and product descriptions in there. I had no idea the free tier gave them rights to use my content like that. Has anyone else dug into their tool's privacy policy and found something sketchy?

3 comments

3 Comments

jakewhite1mo ago

And that's the thing that gets me, it's not just the free apps anymore. Some of the paid tools are sneaking in these clauses too, burying them in sections nobody reads. I've heard stories about people's entire business proposals ending up in training data because they checked the wrong box. It's like these companies know most of us will just click "agree" without reading 20 pages of legalese. The real kicker is once your stuff is in that dataset, there's no getting it back. You can delete your account, email the support team, whatever, but those text snippets are floating around forever. Makes you wonder how many competitors are getting free access to each other's work through these backdoor data collections.

evan_cooper731mo ago

It's like every free app is just data farming in disguise now.

sean8541mo ago

Did you hear about that meditation app that got caught uploading people's private journal entries to train their chatbot? My cousin used it for six months before he found out.