Taiwan’s AI agent shopping debate reaches regulators as retailers call for three rules

Taiwan’s debate over AI agents that place orders is moving into the regulatory spotlight. According to BigGo Finance, the country’s Non-Store Retailers Association has publicly called for three rules to govern agent-driven shopping, and the Ministry of Digital Affairs has agreed to study a risk framework.

The reporting frames the issue as “AI agent ordering” — software acting on a shopper’s behalf to pick products, compare prices and, in some cases, submit the order across retail and e-commerce channels.

The association’s ask is for three rules that would put agent participation in transactions inside a framework that can be discussed and assigned responsibility, rather than leaving the question to individual platform terms of service.

That agreement to study a risk framework puts the debate at the pre-legislative stage, where the first task is defining the problem rather than writing binding rules.

The questions in play are familiar ones. Did the consumer clearly authorise the purchase? Who is liable when an agent buys the wrong item or overshoots a budget? Do platform terms even allow automated programs to act for a user? None of these are new, but handing part of the decision to a machine sharpens them.

What to watch next is the scope of the framework: which transaction types and platforms it covers, whether cross-border marketplaces fall inside it, and whether the association’s three rules eventually turn into concrete text.