Former Meta Security Chief Calls Password Access for AI Agents 'Insane' (epyhia.com)

Meta has introduced a new feature called "spaces" for its Muse AI assistant, allowing users to create interactive hubs for various tasks, such as shopping dashboards and planners. This innovative capability lets users pull live data from connected services like Gmail and Google Calendar, enabling a more seamless and dynamic interaction. With the addition of "Active Memory" and "Active Focus," Muse aims to enhance its functionality by recalling relevant past information and helping users work toward their goals more effectively. However, the rollout has raised significant security concerns. Alex Stamos, former chief security officer at Meta, has criticized the practice of giving AI agents access to sensitive personal data such as password vaults, labeling it "totally insane." He emphasizes the risk that comes with granting broad permissions to AI agents, particularly regarding financial transactions, which can leave users vulnerable. Stamos' concerns reflect a growing tension between the benefits of AI automation and the potential for misuse, prompting banks—like JPMorgan Chase—to implement restrictions against such practices. As Meta pushes to expand Muse's capabilities, the effectiveness of its security measures remains to be seen, leaving users to weigh the convenience of AI assistance against the risks involved.