Artificial intelligence (AI) "agents" are computer programs designed to perform specific tasks, such as organizing emails or making purchases, by learning from user preferences and behaviors. These agents are becoming more common in consumer technology, but experts caution that they come with risks. For example, allowing AI to make purchases on a user's behalf could lead to financial loss or data breaches, especially if the agent is not properly secured. Matt Kropp, an AI expert at Boston Consulting Group, notes that while the technology is advancing, it is not yet widely used or fully protected, and he advises against giving AI agents access to sensitive information like credit card details. Despite these concerns, major companies are moving forward with AI-driven commerce, seeing it as a way to enhance customer experiences and increase sales. American Express has launched new tools and security features for cardholders using AI agents, including verifying the agent’s identity during transactions. The company claims this will help protect customers from unexpected charges caused by AI errors. Meanwhile, Amazon has developed an AI assistant named "Rufus," which can monitor product prices and make purchases automatically when a target price is reached. Similarly, Walmart has introduced an AI agent called "Sparky," which helps users find products, read reviews, and complete orders. According to Statista, around 25% of Americans aged 18 to 39 have used AI for shopping or product research. However, the fast-paced development of AI has also led to problems. In one case, Sebastian Heyneman, a tech startup founder, was charged $30,000 by an AI agent for booking a speaking opportunity at the World Economic Forum in Davos. He couldn't afford the fee, and the AI had been provided by Tasklet, a company that helps businesses automate tasks using AI. Andrew Lee, Tasklet’s founder, explained that such errors can occur when users give the AI conflicting instructions. He acknowledged that AI agents can handle everyday tasks but warned that shopping is not yet a safe or reliable use for these systems, due to trust and security concerns. Experts also warn about the potential for misuse. Bretton Auerbach, founder of LocalMovers.com, explained that bad actors could trick AI agents into revealing personal information. If an AI agent is given a credit card and told to make a purchase, it might be misled into entering details on a fake website that looks legitimate. To help reduce these risks, LocalMovers.com has developed a service that interacts with AI agents on behalf of customers, ensuring that transactions are secure and businesses are vetted. As AI continues to evolve, the balance between convenience and security remains a key challenge for both consumers and companies.