AI Agents Revolutionize Shopping: Are Risks Worth It?

Artificial intelligence agents are being touted as capable of everything from organizing emails to purchasing items based on personal preferences. However, technology experts caution that entrusting key decisions to AI could lead to communication errors and financial losses, while also potentially providing hackers with access to sensitive data.

“Matt Kropp, an AI expert at Boston Consulting Group, told CBS News that it isn’t mainstream yet and is quite risky because there aren’t enough safeguards for people to feel comfortable with agents making purchases on their behalf. ‘It could potentially go buy a car, but I wouldn’t say, ‘Here’s my credit card.'”

Despite these concerns, some of America’s largest companies are moving forward with AI commerce, seeing it as a new way to engage customers and drive sales by letting the AI do the legwork for shoppers.

For instance, American Express recently announced new services and protections for cardholders who make purchases using specified AI agents. This includes verifying the identity of an agent when making a purchase, according to the credit card issuer. The service will protect eligible customers from charges related to AI agent errors.

Amazon’s agentic AI assistant, Rufus, can track product prices on its platform, alert customers when the price hits a certain level and complete the purchase. Walmart, the largest U.S. retailer, has deployed an AI agent named Sparky that it says can help consumers find products, provide reviews and assist with ordering.

About a quarter of Americans aged 18 to 39 have tried using AI for product research or shopping, according to November data from market research firm Statista.

The rapid adoption of AI is also leading to mishaps. For example, Sebastian Heyneman, the founder of a San Francisco-based tech startup, instructed an AI agent to secure him a speaking opportunity at the World Economic Forum in Davos, Switzerland. The bot succeeded but charged him $30,000 for the privilege.

Andrew Lee, the founder of Tasklet, told CBS News that such problems can arise when a user gives conflicting instructions to the AI. He also said agentic AI is fully capable of shopping for people and doing “normal things consumers can do.” But he warned against using these systems for shopping because they are hard to trust.

“Personally, I am not super comfortable with that yet,” Lee told CBS News. “As a business, we don’t recommend it.”

The reason: Bad actors could trick AI agents into handing over personal information by making them believe they’re visiting a legitimate website when in fact they’re on a phishing site.