The Fragility of Reputation under Manipulated Information

Reputation is a central mechanism for sustaining cooperation in human societies, yet–unlike in typical theories and most experimental setups–real-world reputational information about others’ past behavior is often noisy, delayed, or deliberately manipulated, raising the question of whether cooperation can survive when reputational signals are unreliable. We experimentally study how cooperation evolves when individuals can strategically manipulate their public image and whether allowing others to verify such information can counteract manipulation and restore cooperation. As a benchmark, we replicate a standard repeated network cooperation game in which participants observe one another’s past actions and freely choose partners for future interactions. This baseline produces stable, high levels of cooperation driven by conditional association and reputation-based reciprocity. When participants can manipulate information about their past behavior at a cost, manipulation becomes widespread and cooperation collapses. Introducing the option to verify others’ reputations does not prevent this decline: verification is infrequent, poorly targeted, and insufficient to restore reliable reputational information. Consequently, cooperation disappears even when truth-checking is possible. These findings expose the vulnerability of reputation-based cooperation to strategic distortion and highlight the crucial role of trustworthy, verifiable reputation systems. This fragility is magnified in the modern digital era, where profit-driven communication technologies and social media make distortion, selective presentation, and manipulation of information cheap and widespread. Such dynamics foster the spread of disinformation, including fake news, climate change denial, and vaccine refusal, undermining collective efforts to address global challenges such as climate change, resource depletion, and pandemics.