Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Updates to Consumer Terms and Privacy Policy

anthropic.com

Post reply on HN
Print view View original

251–260 of 551 posts

Re: Updates to Consumer Terms and Privacy Policy

#251

Post by andrewmcwatters »

Syzygies wrote: post #193

Claude assists me in my math research. The scenario that concerns me is that Claude learns unpublished research ideas from me as we chat and code. Claude then suggests these same ideas to someone else, who legitimately believes this is now their work. Clearly commercial accounts use AI to assist in developing intellectual product, and privacy is mandatory. The same can apply to individuals.

A lot of people doing cat-and-mouse threat detection development are keeping their work outside of public LLMs right now, so it sounds like you’re in the same boat as a lot of us.

Re: Updates to Consumer Terms and Privacy Policy

#252

Post by general1726 »

aosaigh wrote: post #122

Does this include any code base you are running Claude Code with (where parts of code are sent as part of the context)? I'm not hugely clear on how my private codebase is exposed to Claude in the first place when using Claude Code.

I would expect that your whole code base would be then used as training data.

I think that it is only a matter of time before they will start reselling these data as exfiltrated IP to whoever will be interested.

Re: Updates to Consumer Terms and Privacy Policy

#253

Post by y-curious »

santiagobasulto wrote: post #228

It's obvious this was going to happen. If you ever used any cloud AI model without acknowledging that at some point your data might be leaked (in one way or another) it's on you, not on them. As some other user put it: "big corp changes policy and breaks promises, how shocking"

At least they didn't drop the "Don't be evil" clause quietly

Re: Updates to Consumer Terms and Privacy Policy

#254

Post by thisOtterBeGood »

Syzygies wrote: post #193

Claude assists me in my math research. The scenario that concerns me is that Claude learns unpublished research ideas from me as we chat and code. Claude then suggests these same ideas to someone else, who legitimately believes this is now their work. Clearly commercial accounts use AI to assist in developing intellectual product, and privacy is mandatory. The same can apply to individuals.

This perfectly describes one of the biggest dillema with AI. Where does an AI company stop to utilize human knowledge it does not actually own. Where do they draw the line. Apparently it's possible there aren't any lines drawn at all.

Re: Updates to Consumer Terms and Privacy Policy

#255

Post by bubblyworld »

the_arun wrote: post #243
psychoslave wrote: post #28

What a surprise, a big corp collected large amount of personal data under some promises, and now reveals actually they will exploit it in completely unrelated manner.

From Anthropic communication: > If you’re an existing user, you have until September 28, 2025 to accept the updated Consumer Terms and make your decision. If you choose to accept the new policies now, they will go into effect immediately. These updates will apply only to new or resumed chats and coding sessions. After September 28, you’ll need to make your selection on the model training setting in order to continue…

Under the FAQ:

> Previous chats with no additional activity will not be used for model training.

Re: Updates to Consumer Terms and Privacy Policy

#256

Post by ravishi »

jsnell wrote: post #231
benterix wrote: post #129

Earlier quoted context omitted.

What kind of guarantee do we have this is true? Meta downloaded copyrighted content and trained their models on it, OpenAI did the same. Uber developed Greyball to cheat the officials and break the law. Tesla deletes accident data and reports to the authorities they don't have it. So forgive me I have zero trust in whatever these companies say.

If it were a lie, why take the PR hit of telling the truth about starting to train on user data but lying about the specifics? It'd be much simpler to just lie about not training on user data at all. If your threat model is to unconditionally not trust the companies, what they're saying is irrelevant. Which is fair enough, you probably should not be using a service you don't trust at all. But there's not much of a di…

It's all PR. Some people won't read the details and just assume it will train on all data. Some people might complain and they tell it was a bug or a minor slip. And moving forward, after a few months, nobody will remember it was ever different. And some might vaguely remember them saying something about it at some point or something like that.

Re: Updates to Consumer Terms and Privacy Policy

#257

Post by jsheard »

perihelions wrote: post #59
jsheard wrote: post #41

Earlier quoted context omitted.

Why single out user data specifically? Most of the data Anthropic and co train on was just scooped up from wherever with zero consent, not even the courtesy of a buried TOS clause, and their users were always implicitly fine with that. Forgive me for not having much sympathy when the users end up reaping what they've sown.

Training on private user interactions is a privacy violation; training on public, published texts is (some argue) an intellectual property violation. They're very different kinds of moral rights.

I wish I could be so optimistic that there is no private information published unintentionally or maliciously on the open web where crawlers can find it.

(and as diggan said, the web isn't the only source they use anyway. who knows what they're buying from data brokers.)

Re: Updates to Consumer Terms and Privacy Policy

#258

Post by cactca »

34679 wrote: post #149

I'd bet this is related to their recent decision to boot people for being "abusive" to Claude. It now seems that was an attempt to keep their training data friendly.

This! Any LLM provider that monitors chat/api history for ‘abuse’ towards the model is considering using user data for training.

An Effective Altruism ethos provides moral/ethical cover for trampling individual privacy and property rights. Consider their recent decision to provide services for military projects.

As others have pointed out, Claude was trained using data expressly forbidden for commercial reuse.

The only feedback Anthropic will heed is financial and the impact must be large enough to destroy their investors willingness to cover the losses. This type of financial feedback can come from three places: termination of a large fraction of their b2b contracts, software devs organizing a persistent mass migration to an open source model for software development. Neither of these are likely to happen in the next 3 months. Finally, a mass filing of data deletion requests from California and EU residents and corporations that repeats every week.

Re: Updates to Consumer Terms and Privacy Policy

#260

Post by Workaccount2 »

latexr wrote: post #194

So far, no one in this thread seems surprised (or is admitting to it). But I genuinely would like to know if someone is surprised. I’d also like to know what lead you to believe this wouldn’t happen and if there’s anyone in the LLM space you’d trust to not pull the same stunt (and why do you still believe that). I genuinely want to know and would like to have a productive conversation. I would like to identify what m…

A lot of people hold Anthropic as the "clean and ethical" AI company. They're not power hungry, they are focused on safety. Their aesthetic is cool modern valley vibes. Well grounded and in touch. They don't have the stench of Altman or the ominous presence of the tech giants. They make claude code which is the darling LLM of pure souled silicon valley.
Post reply on HN