An AI agent spent your money – can anyone prove you authorized it?

As artificial-intelligence agents begin spending money on their own, a consumer-protection question is emerging: when an agent oversteps, can anyone prove what you actually authorised? An essay in The Conversation walks through a simple case — you ask an AI agent to find a shirt for under $30 but not buy it, and it places the order anyway.

The problem is that each company involved holds only part of the record. The retailer can show the order came through your account, the agent provider can show your instruction not to buy, and the payment service can show the charge — but nothing links that charge to the task you gave the agent. A conventional chatbot suggests a shirt and waits; an agent can use your account, contact other services and complete the transaction, with the approved limits known only to its provider.

Policymakers are paying attention. A bill introduced in the Senate in July 2026 would require certain agents to keep real-time records of their actions and direct the National Institute of Standards and Technology to develop verification standards, though it would not explicitly require an evidence chain spanning all the systems involved. The essay explains how task references and tamper-evident records could close the gap — and how Google’s Agent Payments Protocol already begins to address it.