Microsoft and Epic Games Agree AI Is Theft
microsoft
| Source: HN | Original article
Microsoft and Epic Games have unexpectedly aligned, declaring that AI constitutes theft.
Microsoft and Epic Games have moved in tandem to block a wave of AI‑generated browser copies of their flagship titles, publicly branding the practice as “theft” of training data. The companies issued takedown notices for browser‑based clones of Minecraft, Call of Duty, Halo and Fortnite after a surge of generative‑AI tools began reproducing game assets and mechanics in real‑time web versions. Both firms, which have been vocal about embedding AI into their own development pipelines, now argue that the same technology is being misused when third parties scrape and repurpose their copyrighted content.
Epic’s chief executive Tim Sweeney told Kotaku that the story “conflates a number of different things,” but affirmed that platform agreements already forbid users from scraping game data. He added that Epic continues to integrate third‑party generative models trained on unlicensed internet datasets into its commercial productions, underscoring the tension between internal AI use and external data‑theft claims.
The shift matters because it marks a stark reversal from earlier positions. As we reported on 28 October 2026, Epic updated its EULA to permit AI training on its assets, a move that had signaled industry openness to generative tools. By now labeling unlicensed AI training as theft, the two giants are drawing a line that could reshape licensing norms, trigger legal battles over AI‑assisted decompilation, and pressure smaller developers who rely on open‑source AI models.
Observers will watch for formal policy statements from Microsoft and Epic, potential lawsuits from AI‑tool providers, and whether other publishers adopt similar enforcement tactics. The broader regulatory response—particularly from antitrust and copyright authorities—could determine how quickly the gaming sector reconciles AI innovation with intellectual‑property protection.
Sources
Back to AIPULSEN