Google AI's Project Genie generates Nintendo game clones, raising copyright concerns

Google AI Launches Project Genie for Interactive World Creation

Google has begun rolling out access to Project Genie, an experimental AI tool that allows Google AI Ultra subscribers in the United States to create playable worlds from text prompts1

1

. Powered by Genie 3 and Gemini, the technology generates interactive environments in real-time as users explore them, with no pre-loaded map in memory2

2

. The tool first uses Google's image generator Nano Banana Pro to create a preview, which users can adjust before the Genie 3 world model generates the environment2

2

. Users can move through these worlds by walking, running, swimming, or flying, though experiences are currently limited to 60 seconds at 720p resolution and 24fps3

3

.

Nintendo Games Clones Emerge Immediately

The Verge's Jay Peters, granted early access to the AI tool, demonstrated Project Genie's ability to generate playable reproductions of Nintendo games including Super Mario 64 and The Legend of Zelda: Breath of the Wild1

1

. The Breath of the Wild clone featured a Link-like character running through a Hyrule-inspired landscape, accurately deploying a paraglider when jumping off cliffs3

3

. Peters also generated worlds resembling Metroid Prime 3 and Kingdom Hearts. Other users posted clips showing worlds styled after Harry Potter and Grand Theft Auto2

2

.

Copyright Infringement Concerns Trigger Rapid Blocks

The technology launched with minimal restrictions around copyrighted material, and the model appears to have been trained on intellectual property from major gaming franchises3

3

. Project Genie subsequently blocked Peters from creating additional Super Mario 64 reproductions, displaying a warning message citing concerns for "the interests of third-party content providers"1

1

. Google Deepmind product manager Diego Rivas stated that "Project Genie is an experimental research prototype designed to follow prompts a user provides," adding that "as with all experiments, we are monitoring closely and listening to user feedback".

Limited Game Development Capabilities Raise Questions

While Google describes these as playable worlds, critics note a significant gap between environment generation and actual game development2

2

. The generated worlds offer little beyond basic movement mechanics, with one commenter noting the potential for "vapid, empty open worlds with gameplay equivalent to Superman 64"2

2

. The 60-second limitation and technical constraints of 720p at 24fps further restrict practical applications2

2

. Access requires a Google AI Ultra subscription at $124.99 per month for a three-month package, available only to users over 18 in the United States2

2

.

Echoes of OpenAI Sora Controversy

The rollout mirrors OpenAI's approach with Sora 2, which initially allowed users to generate videos featuring licensed characters from Nintendo and Disney before facing legal pressure. Following that controversy in October, OpenAI Sora implemented restrictions, with CEO Sam Altman claiming surprise that brands objected to unauthorized use of their properties2

2

. Two months later, Disney invested $1 billion into OpenAI to officially license 200 characters for the AI model, describing it as a way to "thoughtfully and responsibly extend" its storytelling. Whether Google pursues similar licensing deals with Nintendo and other game publishers remains to be seen, though the pattern suggests this may be an intentional strategy to attract attention and broker partnerships.