Summary of Key Points
AI payment is becoming a new battleground for tech giants: Alipay has already achieved 300 million commercial AI payments, with companies like JD.com, WeChat, and UnionPay also making significant moves in this area. While AI can assist users in placing orders and making payments, it can even make independent decisions on spending within preset limits. However, this comes with various trust issues (such as AI-induced errors, default enabled permissions, and blurred responsibility), concerns about platform control (fear of being “pipelined” by AI agents), and the risk that “seamless payment” could lead to “unnoticed debt.” Behind the convenience lies users’ worries about “trust” and “choice.”
What is AI Payment?
In simple terms, AI payment allows AI to handle the entire process of placing an order and making a payment on your behalf, without you needing to navigate to a payment page. For example, with Alipay’s “AI Pay,” you can say “Buy me a cup of milk tea” in the chat window, and the AI will select a store, place the order, and make the payment for you. JD.com’s ClawTip goes a step further; it allows the AI to compare prices and decide what to buy within a set limit (e.g., 500 yuan per month), essentially giving the AI an “allowance account.”
The core of these functions is the “AI agent,” which understands your needs and automatically utilizes various services (such as delivery platforms and payment channels) to complete the transaction. Not only Alipay and JD.com but also WeChat Pay, which has integrated with Tencent’s Meta Platforms, and UnionPay have introduced AI-based payment solutions. Even international companies like Visa and PayPal are exploring “agent-based commerce,” all aiming to make AI the primary channel for users to spend money.
Why Are the Giants Competing for AI Payment?
This is because it represents the ticket to the next generation of user interaction. Over the past decade or so, the payment industry has undergone a phase of “pipelining”: users no longer use bank apps directly but link their cards to platforms like Alipay and WeChat, with banks handling the settlement in the background. AI agents could lead to another round of “pipelining.” In the future, you might simply ask an AI to book a hotel and flight for the weekend, and it will handle all the details without you knowing which payment service is being used.
Whoever controls the AI agents will control the users’ ability to spend money. Companies like Alipay and WeChat are eager to avoid becoming marginalized, just as banks did in the past. Platforms with fewer consumer-centric use cases, such as DuXiaoman, see AI payment as an opportunity to gain traction—after all, users don’t frequently use them, but if AI can make spending decisions on their behalf, they could reach them more effectively.
Trust Issues with AI Payment
The biggest challenges with AI payment include uncertainty:
- AI-induced errors: For instance, some users have experienced buying insurance through AI only to find out the AI mistakenly used a personal payment code, resulting in a loss of 1618 yuan due to a “large model illusion.”
- Default-enabled permissions: Users may disable AI payment features but still find them reactivated later. A previous consent for a milk tea discount could turn into a permanent authorization.
- Responsibility: Alipay’s AI wallet agreement states that AI actions are considered the user’s own, meaning users bear the risk of any mistakes.
Ant Group has acknowledged that the lack of widespread adoption of AI payment is not due to technical limitations but rather users’ reluctance to trust it. After all, payment involves a trust in the system, and AI is particularly prone to causing unexpected issues.
DuXiaoman’s Struggles
DuXiaoman, backed by Baidu, lacks prominent consumer scenarios and has struggled to gain user trust. It relies heavily on advertising (targeting popular shows and dramas) but faces many complaints, including about aggressive debt collection and harassment. In 2024, it was fined 1.21 million yuan by the central bank for violations.
For DuXiaoman, AI payment presents an opportunity to transform its business model. If it can use AI to make financial decisions on users’ behalf, it could move away from relying on traditional traffic sources and become a full-fledged AI finance company. However, users may question the trustworthiness of a platform that hasn’t even resolved basic trust issues in traditional finance.
The Risk of “Seamless Payment” Leading to “Unnoticed Debt”
The trend towards more seamless payment methods (from entering card passwords to password-free transactions) is ongoing. If AI not only makes payments but also makes financial decisions, users might find themselves in debt without realizing it. Financial activities should involve clear risk awareness, and if all processes are automated in the background, users could easily incur debts without knowing it.
The Boundary of Convenience
The trend towards AI payment is unstoppable, but it must not come at the cost of users’ awareness and control over their financial choices. Technology needs to address potential errors, regulations must clarify responsibilities, and platforms must ensure that users can freely enable or disable these features without affecting essential services. The next step in seamless payment shouldn’t be the creation of unnoticed debt.
In summary, while AI payment offers convenience, it’s crucial to balance it with user trust and control over their financial decisions.