No-AI clauses are becoming increasingly common in videogame contracts as developers and publishers grow more concerned about the legal and commercial risks surrounding generative AI.

Haley MacLean, a corporate intellectual property lawyer and videogame specialist at Voyer Law, told GamesRadar+ that these clauses have become especially common over the past year.

According to MacLean, no-AI provisions were relatively unusual several years ago. That has changed considerably, with even smaller publishers now asking for language that prevents generative AI from being used during development and other parts of a game's production.

MacLean said that if a contract does not include restrictions on generative AI, she would push for them, particularly when the publisher is also responsible for creating assets or handling areas such as marketing, porting, and quality assurance.

The shift reflects growing concern across the industry about how generative AI affects both game development and intellectual property.

Generative AI has been promoted as a way for studios to create artwork, music, dialogue, and other content more quickly. At the same time, the technology has faced criticism over the copyrighted material used to train many AI systems and uncertainty surrounding ownership of AI-generated work.

That uncertainty can create serious problems for game developers.

One of the biggest issues is that simply generating an asset with AI does not necessarily give a company clear ownership rights over the finished material. If an asset has not received enough human creative input, it may be difficult to establish copyright protection over it.

MacLean pointed to AI-generated environmental assets as an example. If a studio used generative AI to create hundreds of trees for a game and did not have artists meaningfully transform those assets, the studio could potentially struggle to claim exclusive ownership over the resulting designs.

That creates an uncomfortable situation for companies attempting to use AI to reduce development costs.

A studio may use AI to generate an initial asset, but if it wants stronger legal protection, human artists may still need to substantially modify and develop that material. That can reduce some of the cost savings companies expect from replacing parts of the creative process with AI.

The risks are not limited to copyright.

Developers also have to consider how players respond to the use of generative AI. Recent research has suggested that games disclosing AI use can receive fewer reviews and more negative reactions from players.

That has contributed to a growing concern among developers that using generative AI could damage their relationship with their audience.

MacLean said she increasingly hears the sentiment that developers do not want to use generative AI because their communities would react negatively.

This leaves publishers and developers facing several competing pressures. Companies want to reduce costs and increase efficiency, while developers have to consider copyright, ownership, player perception, and the potential for future legal disputes.

MacLean expects those disputes to become more common as companies continue experimenting with the technology.

She believes the industry is currently caught between the excitement surrounding generative AI and a legal system that is still working to determine how the technology should be treated.

For game studios, that uncertainty is already showing up in contracts. What was once an unusual provision is increasingly becoming standard language designed to protect developers from unwanted AI use and the legal problems that could follow.

As generative AI becomes more common in game development, those clauses could become an increasingly important part of how studios manage their relationships with publishers.