Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Mecha-nudges for Machines

Giulio Frey, Kawin Ethayarajh

Abstract

Nudges are subtle changes to the way choices are presented to human decision-makers (e.g., opt-in vs. opt-out by default) that shift behavior without restricting options or changing incentives. As AI agents increasingly make decisions in the same environments as humans, the presentation of choices may be optimized for machines as well as people. We introduce mecha-nudges: changes to how choices are presented that systematically influence AI agents without degrading the decision environment for humans. To formalize mecha-nudges, we combine the Bayesian persuasion framework with V-usable information, a generalization of Shannon information that is observer-relative. This yields a common scale (bits of usable information) for comparing a wide range of interventions, contexts, and models. Applying our framework to product listings on Etsy -- a global marketplace for independent sellers -- we find that following ChatGPT's release, listings have significantly more machine-usable information about product selection, consistent with systematic mecha-nudging.

Mecha-nudges for Machines

Abstract

Nudges are subtle changes to the way choices are presented to human decision-makers (e.g., opt-in vs. opt-out by default) that shift behavior without restricting options or changing incentives. As AI agents increasingly make decisions in the same environments as humans, the presentation of choices may be optimized for machines as well as people. We introduce mecha-nudges: changes to how choices are presented that systematically influence AI agents without degrading the decision environment for humans. To formalize mecha-nudges, we combine the Bayesian persuasion framework with V-usable information, a generalization of Shannon information that is observer-relative. This yields a common scale (bits of usable information) for comparing a wide range of interventions, contexts, and models. Applying our framework to product listings on Etsy -- a global marketplace for independent sellers -- we find that following ChatGPT's release, listings have significantly more machine-usable information about product selection, consistent with systematic mecha-nudging.
Paper Structure (34 sections, 1 theorem, 20 equations, 5 figures, 11 tables)

This paper contains 34 sections, 1 theorem, 20 equations, 5 figures, 11 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Nudges
  4. Bayesian Persuasion
  5. Usable Information
  6. Formalizing Mecha-nudges
  7. Systematic Evidence of Mecha-nudging
  8. Data
  9. Methods
  10. Results
  11. Machine-usable information rises significantly after the release of ChatGPT.
  12. This result is robust to a wide range of possible confounders.
  13. The effect cannot be explained by AI-assisted copywriting.
  14. Token-level patterns offer a partial window into the mechanism.
  15. Limitations & Future Work
  16. ...and 19 more sections

Key Result

Proposition 1

Consider a bounded-receiver analog of Bayesian persuasion in which both the choice architect and decision-maker have log-scoring utility $\log_2(\cdot)$, and the decision-maker is restricted to predictive family $\mathcal{M}$. Then $\mathop{\mathrm{arg\,max}}\limits_{\tau \in \mathcal{T}} I_{\mathca

Figures (5)

  • Figure 1: After the release of ChatGPT in Nov 2022, the change in machine-usable information in Etsy listings increases significantly, from $\sim 0$ to $0.143$ bits. The change relative to the Jul-Oct 2022 period is plotted here. The effect attenuates over the following year before climbing again in late 2024. This coincides with the release of ChatGPT Search, which could browse live listings---unlike earlier models that only could surface listings from their training data.
  • Figure 2: Our pipeline for estimating the change in usable information between the pre- and post-ChatGPT periods: generate buying decision labels, train content and null models for each period, and run an OLS regression of the pointwise $\mathcal{V}$-information (pvi). This describes our baseline experiment, of which we run many variations (including with controls).
  • Figure 3: The increase in machine-usable information post-ChatGPT is robust to possible confounders: a generic temporal change (DailyMed), AI-assisted copywriting (Rephrase), controls for product- and seller-specific attributes (green), the model family that is fine-tuned to estimate pvi (red), and the LLM used to generate training labels (purple), among others (Appendix \ref{['appendix:label_construction']}, \ref{['appendix:controls']}). Unless otherwise specified, we use GPT-5-mini as the labeling model and Llama-3.1-8B as the fine-tuning model. Each point reports the OLS estimate of the post-ChatGPT shift in pvi, with 95% confidence intervals.
  • Figure 4: (left) Words with among the largest impact on how much machine-usable information Etsy listings have (see Table \ref{['tab:token_ablation_n25']} for a full list). A positive $\Delta$pvi means that the word, on average, makes the machine behave more predictably; negative $\Delta$pvi means that, on average, it makes the machine behave less predictably. (right) Etsy product categories where human buyers are ostensibly sensitive to AI usage (e.g., art and collectibles) show no mecha-nudging (based here on Gemma-3-27B labels).
  • Figure 5: Distribution of pvi scores across quantile-spaced bins for listings uploaded before and after the ChatGPT release (November 30, 2022), using SELECT/PASS tokens with balanced sampling. Each panel corresponds to a different labeling model. The $y$-axis reports the fraction of observations in each bin.

Theorems & Definitions (6)

  • Definition 2.1: $\mathcal{V}$-Usable Information
  • Definition 2.2: Pointwise $\mathcal{V}$-Information
  • Definition 3.1: Mecha-nudging Design
  • Proposition 1: Bounded-Receiver Bayesian Persuasion
  • Definition 3.2: Realized Mecha-nudge
  • proof