The Globe and Mail reports in its Thursday edition that On Oct. 21 Fouad Matin, head of OpenAI's cyber safety efforts, was running on two hours of sleep while addressing two critical security vulnerabilities discovered by artificial intelligence. A New York Times dispatch to The Globe reports that this work involved security researchers from start-ups, tech giants and government officials. Mr. Matin says AI "really opens this Pandora's box of vulnerabilities." AI companies have a responsibility to track down and fix problems, he adds, as "the true goal is to scan and patch as much of the world's software as possible." In Silicon Valley, technologists like Mr. Matin are facing a major shift in cybersecurity. AI is uncovering software flaws rapidly, leading to a surge of vulnerabilities that engineers are struggling to address before hackers exploit them. Companies are increasing their resources to tackle these issues, causing some technologists to panic. OpenAI's technology started scanning for flaws in open-source software in May, identifying over 1,300 bugs through its models, utilized by both internal and external researchers. In June, Google also used AI to find and fix 1,072 bugs in its web browser Chrome.