Sony paid a $50,000 bounty to Jordy, a modder who privately submitted a critical PS5 vulnerability. The exploit reportedly allows users to run Linux on the PS5, with research covering firmware versions 13.60 to 14.0.
Jordy stated that he used large language models to help locate the issue. However, another security engineer had independently discovered the same vulnerability earlier and was researching how to run Linux on PS5 consoles with the affected firmware. Thus, this wasn't the first time an LLM created a entirely new attack method; rather, such tools lowered the barrier for other researchers to find the same vulnerability.
Jordy believed that even if he didn't report it to Sony, someone else would likely discover the issue using similar tools soon. He ultimately decided to submit the vulnerability and claim the bounty, a choice also influenced by the financial pressure from his mother's terminal illness.
After Sony patched the vulnerability, the original plan to run Linux on the console while retaining normal gaming capabilities will no longer work. This incident shows that generative AI may impact console security research and the jailbreak ecosystem, but this single case is not enough to prove that jailbreaking has become easier than ever before.