James Pond’s New Trailer Has AI Alarm Bells Ringing, But Dev Says Otherwise - GameSpot
James Pond has ventured into the world of generative AI before. As for the collection of James Pond games? The compilation includes James Pond: Underwater Agent, James Pond 2: Robocod, James Pond 3: Operation Starfish, and The Aquatic Games Starring James Pond and the Aquabats.
GameSpot may receive revenue from affiliate and advertising partnerships for sharing this content and from purchases through links.
It’s a good time to be a fan of retro video games, thanks to all the various compilations hitting the scene, and for the James Pond enthusiasts out there, their time has finally come. Developer System 3 has announced the James Pond Legacy: The Pond is Not Enough collection, but the reveal trailer is being dragged over claims that it features generative AI art and assets. Comments on the video quickly started pointing out the telltale signs of AI images, prompting System 3 to temporarily shut off public commentary on YouTube. However, the developer insists it didn’t use AI for the trailer.
A cursory glance suggests that System 3 may have used generative AI to create the key art and character renders for the reveal. Each James Pond character has that glossy, wide-eyed aesthetic that’s commonplace in AI image generation, and each character also appears to have been run through several yellow filters, which is often the result of preference tuning. AI models like ChatGPT and DALL-E default to a “golden hour” color palette that is designed to look pleasing to the eye, as these large language models have been trained on historical art, old photographs, and stylized illustrations that feature yellow-hued lighting.
GameSpot reached out to System 3 for comment on this, and we received the following statement, flatly denying generative AI was used for the characters:
“System 3 has confirmed with [the PR team] that no generative AI was used to create the character artwork highlighted in the trailer. The Pondverse characters were produced through a human-led process beginning with the authentic James Pond pixel graphics.”
We were also provided with documents further outlining the process, as well as images showing how the characters were taken from their pixel-art origins, redrawn in Illustrator and Photoshop, and eventually rendered in 3D and further adjusted in Photoshop. Comments were later enabled on the trailer, with the developer pointing to a blog post detailing the process.
James Pond has ventured into the world of generative AI before. A legacy sequel in the form of James Pond: Rogue AI was slammed by Pond’s creator, Chris Sorrell, for using the technology. “I hate almost everything they do with a passion, all the more so since they duped me into being part of their shambolic Kickstarter campaign all those years back, and I stupidly allowed my name to become associated with their bottom-feeding enterprise,” Sorrell said in an interview with Time Extension last year.
As for the collection of James Pond games? The compilation includes James Pond: Underwater Agent, James Pond 2: Robocod, James Pond 3: Operation Starfish, and The Aquatic Games Starring James Pond and the Aquabats. It’ll launch exclusively on Switch 2 in October 2026. It’s worth noting that the collection also features an interactive historical archive showcasing the original handcrafted artwork from the classic games. These illustrations also look much better in comparison, but it looks like they’re joined by a few more of the possibly-not AI-generated images in the preview screens.
Generative AI remains a controversial topic in video game development, especially when used to create art assets. Multiple games have been caught using the technology, with companies regularly claiming that it was only to create temporary assets before it could be replaced by human-made art–like in the case of Pearl Abyss and Sandfall Interactive.
In one notable example, Larian CEO Swen Vincke (Baldur’s Gate 3) said that the studio used generative AI tools for all kinds of tasks, like exploring new game ideas, developing concept art, and creating placeholder text. After a quick retraction and explanation, Vincke confirmed that the studio was looking to grow its pool of artists. Meanwhile, other studios have taken a strong stand against the use of generative AI, with Slay the Spire 2 developer Mega Crit preferring to use “sh**” art over anything generated by machines when conceptualizing a game.
Got a news tip or want to contact us directly? Email news@gamespot.com
GameSpot may get a commission from retail offers.