Independent security research teams using frontier artificial intelligence models frequently uncover identical zero-day software vulnerabilities simultaneously