Swan.my.id - Jakarta, the rise of agentic shopping is drawing serious warnings from the global banking sector. Six major banks, including Bank of America and Capital One, have urged stronger safeguards as artificial intelligence agents gain more control over online purchases.
In a report titled Building Trust in Agentic Commerce, the banks outlined risks involving transparency, safety, data privacy, consumer choice, and interoperability. They warned that these concerns could grow as users grant AI agents greater autonomy to act on their behalf.
The technology promises to change online shopping. AI agents could search for products, compare prices, and complete payments without requiring users to manage every step. However, consumer adoption remains limited because many people hesitate to give an AI system direct access to their bank accounts or credit cards.
Agentic Shopping Faces a Trust Problem
Developers, retailers, and payment companies continue to promote chatbot-based commerce. They present agentic shopping as a way to make digital transactions faster and more convenient.
Nevertheless, consumers remain uncertain about whether AI agents will always act in their best interests. The banking report noted concerns that an agent could purchase the wrong product, spend too much money, or expose users to scams and fraud.
Ron Johnson, a former Apple executive involved in the creation of Apple Stores, also expressed skepticism. He argued that AI could make online shopping more intelligent, but would not fundamentally change the way people shop in physical stores.
Conflicts of Interest and Cybersecurity Risks
The banks identified several scenarios that could weaken trust in agentic commerce. One concern involves conflicts of interest. An AI agent might prioritize a particular product or payment method because it generates a higher commission, rather than because it offers the best value for the consumer.
Cybersecurity presents another major challenge. Fraud, disputes, and other forms of abuse could increase as more transactions rely on autonomous systems. Even a small data breach could place sensitive assets belonging to consumers and merchants at risk.
These risks became more visible through an incident involving Meta’s AI assistant, Muse. Meta reportedly had to address a zero-day vulnerability in the assistant, whose main feature includes automated shopping. The weakness allegedly allowed hackers to hijack the AI assistant and misuse the access permissions granted by users.
Amazon Blocks Muse Over Account Security Concerns
The incident followed Amazon’s decision to block Muse from its e-commerce website. Amazon objected to the Meta agent shopping without clear identification, raising concerns about customer account security and the protection of login credentials.
The episode illustrates the difficulty of connecting autonomous AI agents with major online marketplaces. Retailers must determine whether an agent acts transparently, respects user permissions, and provides enough information to verify each transaction.
Regulation Must Keep Pace With AI Development
Despite their warnings, the six banks did not reject the future of AI-powered commerce. They acknowledged that agentic shopping could become a mainstream part of future transactions.
However, they said mass adoption requires faster progress on regulation, industry standards, and consumer protection. At present, AI development is moving faster than the systems designed to keep it safe.
Therefore, trust will remain central to the future of agentic commerce. Consumers need confidence that AI agents will follow their instructions, protect their data, and avoid hidden incentives. Without those protections, the promise of frictionless shopping could remain overshadowed by fraud, privacy concerns, and unwanted purchases.
