Should you trust personal AI agents? What security experts say.
The more access you give personal AI agents, the more capable they can be. But should you trust them?
Personal AI agents have become increasingly popular, promising to save users time and money by handling tasks like reading emails, ordering items, and paying bills. However, security experts are raising concerns about the safety and privacy of these tools.
Hed Kovetz, a cybersecurity expert, loves the idea of having a personal AI agent to handle some of his administrative tasks. He would love the convenience, but he doesn't trust commercially available agents like Meta's Muse or OpenAI's Dots with access to sensitive information like email accounts, banking details, and health records. Kovetz believes that these AI agents could be taken advantage of or might go beyond the scope of their given tasks.
AI security experts advise consumers to be cautious about sharing personal data with commercial AI agents, as their security and privacy measures are not yet sufficient. Consumer protection expert Lindsay Owens warns that AI agents could take actions that users never explicitly authorized, leaving the user with limited recourse. She notes that it may be difficult to determine whether an AI agent is working for the user or a retailer.
Experts also warn against giving AI agents full access to personal credit card information, as criminals could potentially manipulate the agents using "prompt injection" techniques. These attacks involve tricking an AI agent into ignoring its original instructions, potentially leading to unauthorized spending or the disclosure of private information.
Consumer education expert Lynette Owens advises against entrusting personal AI agents with too much access and information. While she sees value in using AI agents for online shopping research and deal comparisons, she cautions against letting them handle high-ticket purchases or transactions involving sensitive financial information.
Entrepreneur Chris Franco, founder of Woodridge Growth, takes a different approach. He acknowledges that tech companies have been collecting personal data for years for various purposes, and with AI agents, users can expect even greater benefits. However, Franco still advises giving AI agents strict instructions and closely monitoring their behavior to minimize potential risks.
Written by urgent.news from CBS News's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
