Apple playing catch-up with AI-enabled security flaw hunters

CNBC's MacKenzie Sigalos reports on Apple's race to patch Mac security flaws, where AI accelerates both vulnerability discovery by researchers and the company's own response. The story highlights that hackers used OpenAI and Anthropic models to reverse-engineer Apple's patches and build new exploits in just four hours, while Apple employs similar AI tools to sift through a growing volume of bug reports. The increasing popularity of Mac Minis for always-on AI agents amplifies the remote attack surface. - Apple issued two out-of-cycle MacOS patches in ten days to fix remote access vulnerabilities. - Researchers used OpenAI and Anthropic models to reverse-engineer patches and create a working exploit in four hours. - Apple is using AI tools from Anthropic and OpenAI to manage the rising volume of inbound security reports. - Mac Minis are becoming popular for running always-on AI agents, making remote setups more common. - One researcher found around 40,000 Macs with screen sharing accessible over the internet.