← Series One

The Games We Play

How incentives change behavior, and why rational choices don't always add up.

A decision makes sense relative to the choices other people can make. A firm may rush to deploy because it expects rivals to rush too. A supplier may invest in looking credible when customers have little time to verify claims. This strand asks who can move, what each player can see, what they are rewarded for, and how the answer changes when the same players meet again.

Why it matters

Plans often assume that customers, colleagues or competitors will behave as they did before the rules changed. Faster deals, cheaper signals and autonomous agents can change the payoffs before anyone rewrites the policy. Looking at those responses early makes it easier to find the rules, commitments or feedback that could make a better outcome stable.

Published essays

No essays in this perspective have been published yet.

Planned stories

  • The Race Nobody Can Afford to Win

    Firms rush to deploy autonomous AI because competitors do, even when a more measured approach could benefit everyone. Individually rational decisions can produce collectively harmful outcomes.

  • The Game Behind the Rules

    The important question may not be whether an agent obeys instructions, but what behaviors its incentives reward. Perfect rule-following can yield perverse results.

  • Trust Has a Half-Life

    Repeated interaction makes cooperation possible. Copyable, disposable or reconfigurable AI agents may weaken existing assumptions about identity, reputation and trust.

  • The Market for Convincing Lies

    When sophisticated presentation and persuasive claims become cheap to produce, which signals of quality, competence and integrity will remain credible?

  • When Every Agent Agrees

    Similar AI agents optimizing similar objectives may lead firms into herd behavior, unstable coordination and correlated failures.

  • The Cost of Being Predictable

    Consistency can support good governance, but predictable policies can be strategically exploited. How should AI be designed when others adapt to its behavior?

  • The Commons That Feeds the Machine

    AI consumes knowledge created by people. Could the economics of synthetic answers erode incentives to create and maintain original knowledge?

  • The Negotiation Nobody Attends

    When agents negotiate with other agents, whose incentives prevail and how do humans retain meaningful authority over commitments?

Other perspectives