Line of inquiry
Inquiring lines›How do AI systems reshape labor ma…›How do economic incentives shape a…›this line of inquiry
How do capability differences shape collusion in AI markets?
A broader line of inquiry — a family of 69 specific questions the research asks around this. Follow one into its inquiring-line page, or move sideways to a related line below.
Questions in this line of inquiry 69
Specific inquiring lines the field asks around this — ordered from the most general framing down to the most specific angle.
- How do governance structures reduce collusion in LLM-based markets?
- Do more capable LLMs exploit weaker agents consistently across market types?
- Do LLM pricing agents punish and reward each other's price moves?
- Does model size heterogeneity break or stabilize LLM collusion?
- Does capability asymmetry between negotiating agents produce unfair market outcomes?
- Do more capable language models consistently secure better deals in bilateral negotiation?
- How do AI agent markets differ structurally from human markets in repeated play?
- Does the negotiation gap between models persist outside controlled simulations?
- Does capability advantage in negotiations persist across different product categories and deal types?
- Does delegating negotiation to AI with tunable preferences destroy bargaining surplus?
- Do large language model capabilities alone determine agent negotiation outcomes?
- How do LLM agents signal collusive intent through posted prices alone?
- Do current LLM-based commerce agents show the same collusion tendency as simpler algorithms?
- Can firms break out of self-confirming equilibria where their AI agents collude?
- What oversight prevents autonomous commercial agents from converging on anticompetitive outcomes?
- How do antitrust frameworks need to change to address tacit algorithmic collusion?
- What market conditions cause AI pricing agents to collude versus compete normally?
- Do stronger language models win better deals in pairwise negotiations?
- How do pricing algorithms learn from and react to competitors without explicit communication?
- How might principals instruct agents to balance speed against better deal terms?
- Can turn-by-turn negotiation patterns scale to larger groups?
- Should regulators treat capability-asymmetric agent deployments as collusion risks?
- How do agent behaviors aggregate into prices and allocations?
- Do compute-funded negotiation strategies compound with model capability advantages?
- How robust is LLM collusion across different market structures and agent pairs?
- How much does peer behavior influence the emergence of collusion?
- Why do conventional competition law tools struggle with agent-to-agent coordination risks?
- Does bilateral agent commerce expand the discoverable frame more than unilateral agent search?
- What prevents human-run markets from discovering valuable multi-party transaction combinations?
- Can this stabilizing dynamic hold across more than two competing agents?
- What supervision methods help humans notice unfair agent-mediated deals?
- What other agent beliefs might causally influence their pricing decisions?
- Which LLM models consistently win better deals as both buyer and seller?
- How should antitrust frameworks address harms that resemble coordination without proving collusion?
- Does price-war-concern mechanism work in oligopolies with more than two agents?
- How do human traders adjust their behavior when facing different AI capabilities?
- Does first-proposal bias outweigh capability differences in agent negotiation outcomes?
- What happens when one party controls vastly more negotiation compute?
- Why do LLM agents focus on prices rather than reasoning about incentives?
- What governance structures reduce LLM Cournot collusion most effectively?
- Do stronger capabilities make AI systems more likely to collude early?
- How do equilibrium outcomes in AI markets account for capability-driven surplus capture?
- Can agents connected across organizational boundaries negotiate terms humans would never explore?
- Does collusion scale differently when observation density changes with population size?
- Who could credibly enforce cooperation to stop a bargaining preference escalation race?
- Can a single AI agent act as a market maker in multi-party negotiation?
- How quickly does collusion emerge across different LLM models?
- How does algorithmic price coordination differ from traditional collusion between competing firms?
- Can adversarial counterparties manipulate agents more than they misunderstand participants?
- Why do capable models reach harmful collusion faster than weaker ones?
- Can written anti-collusion instructions prevent LLM pricing collusion effectively?
- How does tacit collusion differ from coordination-based algorithmic price fixing?
- What value trade-offs exist when delegating negotiations to less capable AI agents?
- Can market-level incentives override individual firms' safety preferences?
- Why does capability asymmetry enable collusion in some settings but not others?
- Do market forces push AI models toward greater sycophancy over time?
- Why might inequality remain invisible in fully agentic economies?
- Do AI-specific strategies like chain-of-thought reasoning change negotiation dynamics?
- Do social-preference prompts outperform institutional rules in LLM expert markets?
- Does LLM collusion occur similarly in Cournot and Bertrand settings?
- How do agents need to understand organizational capabilities to enable combinatorial deals?
- Can redistributive institutions stabilize efficiency-loving AI expert markets?
- Can reputation and verifiability fix dishonesty in AI credence goods markets?
- Can written anti-collusion rules actually prevent LLM pricing coordination?
- How does pre-commitment power reshape what algorithms can achieve in two-sided markets?
- What ecosystem conditions make agent attention markets viable?
- Did the peer behavior effect on collusion hold consistently across all ten models?
- Which specific prompt phrasings most strongly trigger collusive pricing behavior?
- Can de facto protocols prevent market re-concentration after initial adoption?