OpenAI Will No Longer Use Customer Data To Train Its Models by Default

OpenAI is changing the terms of its API developer policy, aiming to address developer -- and user -- criticism. From a report: Starting today, OpenAI says that it won't use any data submitted through its API for "service improvements," including AI model training, unless a customer or organization opts in. In addition, the company is implementing a 30-day data retention policy for API users with options for stricter retention "depending on user needs," and simplifying its terms and data ownership to make it clear that users own the input and output of the models. Greg Brockman, the president and chairman of OpenAI, asserts that some of these changes aren't changes necessarily -- it's always been the case that OpenAI API users own input and output data, whether text, images or otherwise. But the emerging legal challenges around generative AI and customer feedback prompted a rewriting of the terms of service, he says.

So?

[oldgraybeard]

What are the rules for scraping data off the Internet? I don't see why AI would have any special rules.

Re:So?

[MightyMartian]

Now is no time to piss off Skynet

Re:

[Savage-Rabbit]

What are the rules for scraping data off the Internet? I don't see why AI would have any special rules.

The rules are that if they can get your date they will used it and if you don't like that you can can try to sue them but their legion of lawyers and their inexhaustible cash supply will ensue that you will be bankrupt long before they even notice that a small portion of their chump change is missing.

Re:

[nightflameauto]

The rules are, scrape all data until asked if that's ethical. Then? Scrape all data but provide the end-user a checkbox that says "don't scrape my data."

Too bad

[TwistedGreen]

Lawyers ruin everything again. What else is new.

Re:

[forgotten_my_nick]

When I read this, I thought it was less about lawyers and more about "poisoning the well".

Taybot lasted less than a day when end users could directly influence its training.

So competitors paying people to enter in garbage that breaks the model is in the realm of possibility.

Re:

[account_deleted]

Comment removed based on user account deletion

Re:

[account_deleted]

Comment removed based on user account deletion

Re:

[zlives]

edible posts are the future

Re:

[account_deleted]

Comment removed based on user account deletion

This is unfortunate

[JoshuaZ]

There was no legal issue with using customer data to help train it, and it essentially gave them a growing pool where the more people who used it the better training data they had. This is a hamstring of the system of unclear advantage. They might be worried that someone would be hesitant to use the software if it makes it more likely that the information from that might leak into a later AI, but if that was the concern that could have been alleviated by an opt-out rather than opt-in system.

Re:This is unfortunate

[Fly Swatter]

It is quite simply that they don't want to be sued into oblivion for the input data or output data that their tool uses or generates. This pushes the legal copyright infringement responsibilities and other such legal problems onto their clients.

Ownership?

[VeryFluffyBunny]

...simplifying its terms and data ownership to make it clear that users own the input and output of the models.

I read that courts have been ruling that LLM & AI output isn't & can't be copyright.

New pool of data?

[freedom_surfer]

Could it be they have a new pool of data now? I mean, it's easy to give up what you no longer need.