The Limits of Verifiability: Credibility and Flexibility in Communication -- by Alessandro Lizzeri, Yichuan Lou, Jacopo Perego

TEXT START: We compare verifiable and unverifiable communication in a sender–receiver setting with partially aligned preferences, where the sender knows more than what her evidence can prove.

The Dissection

The paper formalizes a narrow tradeoff: evidence increases credibility but reduces communicative flexibility when evidence is noisy or incomplete. It studies message design inside a human sender–receiver system, treating the existence of informed human actors as given.

The Core Fallacy

Its strategic blind spot is mistaking better communication for durable human economic relevance. Under the Discontinuity Thesis, AI can automate inference, verification, persuasion, and subsequent action. The paper optimizes the communications protocol while leaving the communicator’s productive necessity unexamined.

Hidden Assumptions

• Human senders retain scarce private information.
• Human receivers remain the relevant decision-makers.
• Evidence production and verification stay institutionally constrained.
• AI does not generate, test, rank, and transmit claims at lower cost.
• Communication value translates into continued human bargaining power.

Social Function

Partial truth with transition-management and prestige-signaling characteristics. The model may be valid within its narrow domain, but generalized socially it becomes an anesthetic: it shifts attention toward optimizing human communication while ignoring the replacement of human communicators.

The Verdict

Technically useful microeconomics, strategically inadequate. It describes a lag-phase friction in which humans still control information channels. Under DT, credibility survives as a function; the human sender’s livelihood does not necessarily survive with it. Ownership and control of the verification-and-action machinery determine who captures the remaining value.