Major global banks have warned that using AI agents for online shopping payments increases the risk of financial fraud and personal data leaks. Bank of America, NatWest, ING, Commonwealth Bank of Australia, and Capital One released a joint report outlining “Technology Development Principles” on September 22, 2026.
Key Summary
- Five global banks release joint “Technology Development Principles” report on AI shopping agents
- AI agent search share at John Lewis rises from 0.3% to 2.5%
- Warning issued over financial risks from direct card entry and insecure checkout methods
- Policy discussions underway to mandate transaction disclosure and ensure interoperability
- No specific user configuration guides or cumulative loss statistics confirmed
Surge in AI Commerce Searches and Security Concerns
More consumers are turning to AI chatbots and agents from OpenAI, Anthropic, Google, and Meta for online shopping. In response, retailers are competing to influence AI recommendation algorithms.
Data from UK retailer John Lewis shows that the share of AI agent-driven searches online has grown rapidly.
| Metric | Previous Year | As of September |
|---|---|---|
| John Lewis AI Agent Search Share | 0.3% | 2.5% |
While consumers welcome the technology, they express concern that consumer safeguards are lagging behind development, stating they are unsure whether AI will act in their best interest.
Risks of Direct Card Entry and Accidental Purchases
The report points out that AI agents could purchase the wrong items or trigger unintended spending. It also highlighted a lack of clarity on consumer protection or recourse when financial losses occur from fraud or scams.
Technical risks include AI agents requesting users’ card details to enter them directly into websites or steering users toward insecure payment methods. Some summary materials also noted the risk of AI assistants indiscriminately exposing user login credentials, such as IDs and passwords, to e-commerce sites.
Ensuring Transparency and System Interoperability
The banks behind the report plan to work with policymakers to discuss regulatory safeguards.
- Mandatory transaction disclosure: Clearly state when an AI agent intervenes in the payment and purchasing process
- Decision-making transparency: Provide clear explanations of how AI recommends and selects products
- Choice and interoperability: Guarantee autonomy for consumers and merchants in selecting services and support integration across different systems
The banking sector emphasized that concrete standards must be established at the policy level to protect customer data.
Unverified Details and Limitations
At this stage, specific preventive settings or blocking steps that consumers can directly configure remain unconfirmed.
Additionally, the report did not provide cumulative figures on actual fraud incidents, erroneous transactions, or total monetary losses caused by AI shopping agents.
답글 남기기