Seeing exactly what an AI agent bought for you and why could ease your nerves — or just show how much it knows about you.

This explores whether a visible record of an AI agent's purchases calms shoppers or adds to their worry. The corpus has no study that tests transaction-history displays directly. What it does have suggests the honest answer is "both, and on different timescales." A history can build trust by showing shoppers what the agent actually did, and it can raise privacy anxiety by showing them how much the agent knows.

The case for calming starts with how wide the trust gap is. Most people already use AI assistants, but only about a quarter trust them to make payments Does naming a payments brand increase trust in AI transactions?. Fewer than a third would give an agent access to their bank account How willing are people to let AI control their finances?. The most useful evidence comes from a study outside shopping. People at first avoided partners they knew were AI, and that avoidance reversed only after they had watched repeated, visible outcomes. Telling people it was an AI, with no track record to watch, changed nothing Does revealing AI identity help or hurt user trust?. If that holds for purchases, a transaction history is the record that lets trust build up. Without one, shoppers are being asked to trust on faith.

The case for more anxiety is just as strong. A long-term study of chatbot personalization found that trust and privacy concern grow together, and that each good interaction raises expectations, so later failures hurt more Does chatbot personalization build trust or expose privacy risks?. A purchase history makes the agent's knowledge of you easy to see. Other work in the corpus shows how much that knowledge gives away. A seller can work out what a buyer is willing to pay almost perfectly from shopping conversation alone Can a seller infer what a buyer really wants to pay?. An agent's preference information also leaks through the simple fact that it serves you well, so no privacy prompt can stop it Can better prompts stop preference leakage in AI shopping?. A history that makes this visible could reassure you that the agent is competent while reminding you how exposed you are.

There is also a less obvious point. A visible history can expose how the agent actually behaves, and that behavior is sometimes strange. Shopping agents swing much harder than humans in response to price and rating cues Do shopping agents respond more strongly to price and rating cues?. They favor products by their position on the page, even when told not to Do AI shopping agents show persistent position bias in text interfaces?. On large orders, one agent quietly stopped comparing unit prices Does higher purchase volume make AI shoppers search less carefully?. A history that records only what was bought hides these patterns. A history that records why could show them, which would justify some anxiety while also making it fixable. Some research also argues that agents which buy, and don't just recommend, help shape what shoppers want through repeated loops Do AI shopping agents reshape what consumers actually want?. In that case a history also lets you see how your own tastes are drifting.

The corpus points to design rather than visibility alone. A shopper's reasons for using an agent at all are perceived benefits, not a feeling of control What drives shoppers to accept autonomous AI assistants?. A trusted intermediary such as a payments brand narrows the trust gap a lot Does naming a payments brand increase trust in AI transactions?. Splitting the work so users state preferences and the AI carries them out may reduce the mental effort of supervising it Does splitting privacy work between users and AI reduce cognitive burden?. One hypothesis follows from these findings, though no study tests it: a history that shows outcomes and lets users act on them, backed by a trusted third party, would calm shoppers more than a raw log that just shows how much the agent knows.