Attacks are moving at machine speed

Attackers are early adopters of AI. Just as businesses are deploying AI to automate workflows and improve decision-making, criminals are aggressively deploying the technology to identify vulnerabilities and exploit them in minutes. Most security teams aren’t built for that speed. This year is on pace to see more than 50,000 published common vulnerabilities and exposures, a new record. At the same time, Booz Allen Hamilton research found that cybercriminals can now go from initial access to fully compromising a system in less than half an hour, sometimes far less. They’re also using AI to write convincing phishing emails and identify targets in minutes, the consulting firm said. The risk goes further than attackers. Agents built to perform everyday tasks can inadvertently expose flaws while doing their work. In one recent case, an agent tasked with booking a gym class found a flaw in the reservation system and used it to bump another gym member out of their spot. Nobody asked it to look for a vulnerability, but it found one anyway. These issues can’t be addressed by hiring more staff or buying additional software tools. Organizations must put AI to work in security to identify thre…