A security vulnerability in ChatGPT’s Mac application could have allowed hackers to grab sensitive user data, according to WIRED. The flaw has since been patched, though the report does not specify when the vulnerability was discovered or fixed.
According to WIRED, the vulnerability highlights an emerging security concern: while much attention has focused on the potential hacking capabilities of AI agents themselves, AI software applications are proving to be attractive and vulnerable targets for attackers. The specific technical details of the vulnerability and the type of sensitive data that could have been accessed were not detailed in the report.
The discovery underscores the dual nature of AI security risks. While researchers and security professionals have been analyzing the offensive capabilities that AI systems might enable for malicious actors, this incident demonstrates that AI applications themselves present security vulnerabilities that need to be addressed. The ChatGPT Mac app joins a growing list of software applications that require ongoing security monitoring and patching to protect user data.