Anthropic takes steps to block Claude from reverse-engineering video games

Anthropic's Claude AI has reportedly begun restricting its capabilities to prevent users from reverse-engineering video game code, according to reports from Tom's Hardware. While there hasn't been an official announcement, the reason seems clear. The gaming community has been making headlines recently, with the GTA6 leaks causing a stir and leading to subpoenas from Take-Two Interactive against Microsoft and Discord.

Additionally, the hacking and modding community have found AI to be a useful tool, making it easier for them to undertake tasks like game decompilation.

This development comes after game decompilers began experimenting with various video game source codes, such as integrating Mario into the Elden Ring world or importing Spider-Man into Batman: Arkham City. Ports of popular games like World of Warcraft, Call of Duty, and Grand Theft Auto: Vice City have also been made playable in a browser, without the consent of the original developers or copyright holders.

The PlayStation 5 jailbreak has further complicated matters, enabling backward compatibility with older PlayStation games and cross-platform play with Xbox-exclusive titles and Nintendo games on a jailbroken PS5.

The potential for lawsuits due to these modifications and ports has prompted AI companies to take action to stop or severely limit these activities. However, this move highlights an ironic irony: frontier AI companies, which initially copied vast amounts of information and art for their training sets, are now concerned with copyright infringement.

Furthermore, this situation could spark an AI arms race driven by unethical incentives, as the least ethical AI might become the most popular if companies take steps to respect copyright law or basic ethics.