Banks Say Consumers Unsure AI Agents Are on Their Side

Six major banks — ASB Bank, Bank of America, Capital One, Commonwealth Bank of Australia, ING, and NatWest — have released a set of principles aimed at establishing guidelines for the development of artificial intelligence (AI) agents in commerce. The banks, which play a crucial role in the payments ecosystem and facilitate millions of payment interactions daily, issued these principles to ensure responsible development of agentic commerce, providing consumers and merchants with choice, control, and flexibility in their transactions while maintaining safety and security.

According to the report, however, consumers are uncertain about whether AI agents will act in their best interests. They worry that AI agents might purchase the wrong item, overspend, or even worse, lose money to scams and fraud. Furthermore, consumers are uncertain about the protection they would receive if things go wrong, and they are unsure of who to turn to in such situations.

The report highlights that agentic commerce carries the potential for increased fraud, scams, and disputes. Some providers might adopt unsafe practices, such as prioritizing payment methods with weaker safeguards or disregarding payment processing standards. Moreover, malicious actors could exploit these agents to carry out fraud or scams, including compromising or impersonating AI agents and merchants, engaging in new forms of social engineering, and wreaking havoc on the overall value chain.

The need for these principles becomes even more critical due to the heightened risks associated with AI agents. In the event of a mishap, there is a lack of clear liability allocation and inefficient dispute resolution processes, as these are not consistently applied across the value chain. Research by PYMNTS Intelligence shows that consumer hesitance exists as they utilize AI agents for shopping purposes.

While 50% of Americans have made a retail purchase aided by AI, only 24% are comfortable allowing an AI agent to shop and pay for them. This discomfort arises from consumers' hesitance to delegate critical aspects such as identity verification, payment choice, and irreversible decisions to AI agents.