In This Story
Microsoft may be leading the generative artificial intelligence race — but that’s not without some stumbles.
Last month, Microsoft announced its Copilot+ PCs with AI features, including the Recall tool, which takes screenshots of users’ activity periodically. The screenshots are encrypted, stored, and analyzed by AI to “understand their context,” serving as a sort of “photographic memory” for users to remember their PC activity, according to Microsoft. But after privacy and security concerns emerged, Microsoft said Friday that the feature will be off by default.
Users will have to enroll in Windows Hello to enable the tool, as well as provide proof of presence to view screenshots in Recall. Microsoft is also “adding additional layers of data protection” such as “just in time” decryption to decrypt Recall screenshots when authenticated users try to access them. Copilot+ PCs are slated to start shipping to customers on June 18.
Kevin Beaumont, a cybersecurity expert who formerly worked with Microsoft, shared a thread on X showing Recall’s potential security flaws. Recall, he found, stored information in a plain-text database, making it possible for attackers to gain access using malware.
Microsoft is “going to deliberately set cybersecurity back a decade & endanger customers by empowering low level criminals,” he wrote. A Microsoft executive addressed criticism by comparing its hurried AI developments to a “journey.”
“We are on a journey to build products and experiences that live up to our company mission to empower people and organizations to achieve more, and are driven by the critical importance of maintaining our customers’ privacy, security and trust,” Pavan Davuluri, corporate vice president of Windows and Devices, said in a statement. “As we always do, we will continue to listen to and learn from our customers, including consumers, developers and enterprises, to evolve our experiences in ways that are meaningful to them.”
Microsoft is reportedly facing scrutiny from the Federal Trade Commission over potential anticompetitive practices in the AI industry. The company is reportedly already under investigation by the FTC over its deal to “acqui-hire” the AI startup Inflection. In March, Microsoft paid Inflection a $650 million “licensing fee” to use its models, and hired most of the startup’s employees — including its chief executive Mustafa Suleyman, who now heads Microsoft’s AI division.