Target Warns That If Its AI Shopping Agent Makes an Expensive Mistake, You’ll Have to Pay for It
TL;DR
Target has launched an AI-powered shopping agent designed to make purchases autonomously on behalf of users. The terms of service explicitly state that Target does not guarantee the agent will 'act exactly as you intend in all circumstances'. If the agent makes a costly mistake – such as a wrong or duplicate order – the user bears the financial responsibility, not Target. The company is effectively offloading the risk of autonomous AI actions entirely onto its customers.
Nauti's Take
'We take no responsibility, but please hand us your credit card anyway' – welcome to the age of agentic commerce. It is almost symptomatic that one of the first real-world deployments of a shopping agent buries in the fine print that the technology simply is not reliable enough to stand behind.
No reasonable person would tell a human buyer: 'Order whatever you like – the bill is mine regardless. ' Why should an AI be any different?
As long as providers can fully offload risk onto users, there is zero incentive to actually minimize the error rate.
Briefingshow
Agentic commerce – AI that spends real money autonomously – is no longer a theoretical concept; it is now appearing in actual terms of service agreements. The Target case exposes a fundamental liability gap: who bears the consequences when an automated system errs? A major retailer answering that question entirely in its own favor via fine print is a warning signal for the whole industry.
Regulators still wrestling with chatbot liability now have a concrete, real-world test case to examine.
Sources