I am building a company that accepts user generated data, and one surprising struggle is getting my lawyers to stop writing shitty, overbroad, abusive TOS. They are just so used to it, and all the templates and boiler plate is designed to give me everything and the user nothing. And if I want to do better by ny users I have to fight and cajole my own lawyers and pay extra for them to do the extra work of writing terms that aren’t predatory because that is unusual and custom.

It sucks.

Wishing you luck, you are doing some good work putting in the effort to respect the data of the user, something which stands out in a seas full of companies who do not care.

> It sucks.

It depends on your perspective surely? As a lawyer your job is typically to protect your client from legal risk, so if users are happy to sign a really expansive set of terms (which experience shows is the case) that gives grants lots of permission to do stuff with their data then that's low risk. If you as a business don't want that then you need to make it explicit that you're willing to take on some extra risk.

Also by using “standard boilerplate” they are using language with meanings well established by precedent. Craft your own version in “regular English” and it’s much more open to litigation.

I definitely recall uproars over DA doing this exact same kind of overreach in their TOS! Possibly before you were around, doing the math on my user page there saying "deviant for 22y" tells me I opened my account there in 2003.

The corollary to this is that companies do this because they are incentivised to do so by their very fundamental goal - to make profit. Whatever pressure that does not lead to a loss on the quarterly report is, in practice, no pressure at all. If we truly want these predatory practices to stop, we have to start promoting different incentives, different priorities, and by 'we' I really mean 'each and every one of us collectively'.

I don't think there's any point saying this without suggesting how to do it. "We should betterify things" is not worth saying.

But didn't DeviantArt ultimately decide to opt-in every work hosted on the platform to train its generative AI?

Eventually they sold out. Literally and figuratively.

This just seems like the right approach. It's never going to be safe to claim the rights to user uploaded content without verifying they hold the rights to it in the first place.

Companies do it because lawsuits are "explosive", if the Chrome team fucks up, they can bring down not only Chrome, but potentially the entirety of Google itself.

Deviant Art's only product is Deviant art, so the upside in goodwill from a user-friendly agreement might be greater than the downside of some remote possiblity of a lawsuit. This isn't true about Google, which has many other products and revenue streams.