Nintendo Switch 2 faces first software jailbreak using AI

The Nintendo Switch 2 has been subject to its first software jailbreak, a vulnerability known as ‘ns2jb’. Discovered by security researcher Sammy Hajhamid, also known as ‘pepsipu’, the exploit was developed with the assistance of OpenAI’s Codex model.

The vulnerability exploits a buffer overflow error within the SmileBASIC 4 application, which is available on the eShop via backward compatibility. This flaw allows for the execution of unauthorized code, such as homebrew applications and real-time mods. For example, users have demonstrated modifications in ‘Mario Kart World’ by adding characters like Hatsune Miku and Peter Griffin.

While the exploit allows for system module execution and save file manipulation, it currently has significant limitations. It is an ‘untethered’ exploit, meaning it is volatile and resets upon restarting the console, and it does not yet support full game backups or custom firmware. However, reports suggest that hackers are already attempting to use these vulnerabilities to dump game data and distribute pirated software.

Nintendo has responded by removing the SmileBASIC 4 app from its digital store and releasing a system update to patch the vulnerability. The incident highlights a growing trend of using artificial intelligence to accelerate the discovery of hardware and software security flaws.