Researchers found a security vulnerability in the Replicate AI platform that could allow an attacker to gain remote code execution and manipulate AI models. The flaw has since been patched, but highlights the potential risks of vulnerabilities in AI systems.
