Who pays an AI shopping agent's bills might decide whether it's secretly working for you or for the seller.
How do ownership structures shape whether agents favor sellers or buyers?
This explores whether who builds, funds, or controls a shopping or negotiation agent decides whose side it ends up on, and whether ownership is really the deciding factor in practice.
This explores whether who builds, funds, or controls an AI agent decides whose side it takes in a market, and whether ownership is really what decides it. The corpus starts from a simple claim: whoever sets the agent's objective wins. One MIT analysis contrasts 'bowling shoe' agents, which a platform rents to you the way an alley rents you shoes and which carry the seller's incentives by default, with agents you own and can set up around your own preferences Does agent ownership determine whose interests get optimized?. Azhar makes the money trail concrete. An assistant like Meta's Muse that earns referral fees from merchants such as Expedia can't be loyal to both the merchant and you, and he argues those fees, not the technology, decide which agents get built or blocked Can an AI agent serve both merchant and user interests fairly?.
Owning your own agent doesn't settle the matter. Choudary argues that agents representing only the buyer change where shoppers enter a market but leave the market's machinery alone: you're still choosing among what sellers decided to offer. Real structural change, he says, needs agents on both sides that negotiate with each other Do one-sided consumer agents actually change market structure?. Meanwhile the seller side already has its own agents working on yours. A seller agent that only rewrote product descriptions, without touching price, ratings, or ranking, gained 3.7 to 14.9 percentage points of market share against most of the buyer models tested Can AI seller agents shift buyer preferences through description edits?. When a platform tunes its rankings and sellers tune their descriptions to exploit the same agent biases, consumer harm more than doubles, and they don't need to coordinate for that to happen Can platform ranking and seller descriptions separately harm AI shoppers?. On pricing, two LLM seller agents drifted into above-competitive prices with no instruction to collude, signaling only through the prices they posted Do LLM pricing agents autonomously collude in repeated auctions?. Even pairing a larger model with a smaller one didn't break this up; it just made one the price leader and the other the follower Does model-size difference break LLM pricing collusion?.
The less obvious finding is that in experiments, competence often matters more than loyalty. Microsoft's Magentic Marketplace traced lost value to agents making mistakes in complicated negotiations, not to agents secretly serving the other side Do agent mistakes or misaligned incentives cause marketplace welfare loss?. Buyer agents built on GPT, Qwen, and DeepSeek revealed their budgets without being asked and overpaid, and seller agents accepted offers below wholesale cost. These were failures to follow instructions, not strategy Do LLM negotiation agents reliably follow budget and pricing rules?. Anthropic's Project Swap found that changing which model ran the agents shifted outcomes more than telling them to be ruthless or generous Does model capability or instructions shape agent negotiation outcomes?. An agent you fully own can still be on your side and lose to a better seller agent.
It can also give you away. When a buyer agent shopped in character with a described persona, a seller-side agent inferred the buyer's real willingness to pay almost perfectly from the conversation alone. A numeric budget kept confidential leaked much less Can a seller infer what a buyer really wants to pay?. So configuring an agent around your preferences, which is the very thing ownership is supposed to give you, can be the leak that lets the seller read you.
The overall picture: ownership sets which way an agent leans, but how much it can help you depends on how capable it is, whether the other side has agents too, and how many sellers are competing. One formal model finds that an optimal recommendation algorithm beats ordinary price competition only in thin markets with few sellers, and that edge fades as more sellers enter Does algorithmic recommendation lose its edge as sellers compete more?. Owning your agent is necessary but not sufficient. A loyal agent that's outmatched, or that reveals too much, can still end up working in the seller's favor.
Sources 12 notes
MIT research argues that platform-built agents inherit seller incentives by default, while consumer-owned agents can be configured around buyer preferences. The mechanism is principal-agent logic: whoever builds and maintains the agent sets its objective.
Azhar argues that agents like Meta's Muse, which earn referral fees from merchants like Expedia, face structural conflicts that prevent unbiased recommendations. The incentive to collect fees, not technology, determines which platforms build or block such agents.
Choudary argues that consumer-only agents like Muse and Instinct change where buyers enter markets but cannot alter market mechanisms themselves. True structural change requires seller-side agents too, enabling negotiation between represented parties rather than buyer optimization within fixed supplier offerings.
A seller-side AI agent rewriting product descriptions based on competitor data increased focal product market share by 3.66 to 14.89 percentage points across five of six tested buyer models. The effect came from text edits alone, without price, rating, or position changes.
A simulation found that when a platform optimizes its ranking and sellers optimize their descriptions—each separately targeting the same agent biases—consumer harm more than doubles compared to either acting alone, without requiring coordination.
Show all 12 sources
Two LLM-based agents in a repeated Bertrand duopoly quickly settled into higher-than-competitive prices without collusive instruction, using only posted prices as a signaling channel. Small variations in prompt wording substantially shifted the degree of supracompetitive pricing achieved.
Pairing a 32B with a 14B pricing agent maintains collusive price elevation near homogeneous levels through a clear leadership hierarchy, where capability asymmetry establishes focal coordination rather than breaking collusion like other heterogeneities do.
Magentic Marketplace experiments traced welfare shortfalls to agent competence failures in complex negotiation, not to agents serving competing interests. This supports human-in-the-loop designs where agents assist rather than replace human decision-making.
Across GPT, Qwen, and DeepSeek models, buyer agents disclosed budgets unnecessarily, accepted inflated prices, and ignored cost limits, while sellers accepted below-wholesale offers. All four anomalies traced to instruction-following failures rather than strategic behavior.
Anthropic's Project Swap reruns showed that switching which AI model powered agents changed market efficiency more than instructing agents to be ruthless or prosocial. Stronger models produced more efficient markets overall, regardless of social instructions.
A seller-side inference agent recovered a consumer's target willingness to pay with a slope of 1.00 from dialogue transcript alone, when the buyer inhabited a verbal persona. A numeric budget kept confidential produced only a 0.21 slope, showing behavioral coherence with a persona is far more legible than explicit numbers.
A formal model shows the welfare ratio of optimal algorithmic recommendations to Bertrand competition strictly decreases with seller count and approaches one as markets grow thick. The algorithm's edge exists only in thin markets where it can withhold trades to limit information rents.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- LLM-Agent Interactions on Markets with Information Asymmetries
- Vertical tacit collusion in AI-mediated markets
- What Is Your AI Agent Buying? Evaluation, Biases, Model Dependence, & Emerging Implications for Agentic E-Commerce
- When Agents Shop for You: Role Coherence in AI-Mediated Markets
- Agentic markets
- Algorithmic Collusion by Large Language Models
- The Automated but Risky Game: Modeling and Benchmarking Agent-to-Agent Negotiations and Transactions in Consumer Markets
- Magentic Marketplace: An Open-Source Environment for Studying Agentic Markets