Online game store Steam won’t host games that use copyright-infringing AI artwork. However, overall AI technology that creates no copyright issues is being encouraged, according to a representative from parent company Valve.
Valve’s goal isn’t to discourage the use of AI on Steam and is instead “working through how to integrate it into [its] already-existing review policies”, the company's PR rep Kaci Boyle told The Verge in an email.
This clarification comes after Reddit user potterharry97 claimed in June Valve is unwilling to publish games with AI-generated content anymore.
The Redditor said Steam rejected a game they submitted because it contained assets created by AI. The platform allegedly sent the user an email saying that since the legal ownership of AI-generated art is unclear it cannot ship such a game. If the developer confirmed they owed all the data used to train the AI that was used to create those assets, the game would be allowed.
Potterharry97 claimed that even after manually manipulating the AI-generated art to make it different, Steam still rejected the game.
The rules and guidelines on Steam already stipulate that “content you don’t own or have adequate rights to” cannot be distributed on its platform.
This issue cuts at the heart of one of the big talking points in the AI bubble: its relationship with copyright laws.
If one were to generalize, there are AI enthusiasts who see value in feeding existing works into models for them to pump out remixed versions of them. In the other camp, you’ve got your working artists and creators who’ve been dealing with industries exploiting them long before generative AI tools became so widely available.
Between April and May 2023, the U.S. Copyright Office organized listening sessions with writers, visual artists, musicians, and filmmakers to hear their views on copyright issues arising from the use of AI. The office is facing challenging questions about the implications of models that train on large amounts of data created by people who haven’t necessarily consented to provide it to companies for free. The U.S. Copyright Office is set to publish the results of these discussions in the coming months.
Last week, a class action lawsuit was filed against ChatGPT creator OpenAI in a San Francisco federal court alleging that the company’s technology trained on texts “copied by OpenAI without consent, without credit, and without compensation.”
Over in the E.U., the European Parliament is pushing for new rules that would require AI models to publish summaries detailing their use of training data protected under copyright law.
