When Privacy Moves ML-Mediated Decisions On Device: Information and Incentive Misalignment in Auctions

Dipankar Sarkar

cs.GTSep 27, 2026

When Privacy Moves ML-Mediated Decisions On Device: Information and Incentive Misalignment in Auctions

Authors: Dipankar Sarkar

Organizations: Skelf Research

Abstract

Moving ML-mediated decision making onto privacy-preserving clients decentralises the economic decision along with the inference. Shared budget constraints then depend on information that cannot be globally current, creating an information-structure failure that conventional pacing is not designed to solve. We study this information misalignment in an auction-logic-faithful on-device simulation with 36 campaigns and 50 devices. Accounting is in dimensionless integer score units; no currency semantics are claimed. Across 30 paired demand paths, proportional Even pacing overspends 17.77% after one tick of staleness and 1,669.31% after 50 ticks under the original 20-times budget pressure. The effect does not depend on that severe a budget: at two-times pressure, 50-tick overspend remains 106.95%. A visible-budget no-sale guard makes zero-lag compliance exact at this score-unit granularity, yet leaves 11.88% overspend at one tick because other devices' debits remain invisible. A declared bursty, heterogeneous-device sweep retains a strictly increasing mean lag curve. We derive a finite-window expected excess-debit bound under conditional charge caps and find positive paired slack in every bounded-value cell. A second, incentive misalignment arises when the ML/pacing score transformation is allowed to change payment units: 98.23% of rival auctions at one tick admit a profitable deviation. An executable implementation-level counterexample isolates the runner-up's multiplier in the winner's price. Critical-base-bid payment is per-auction DSIC conditional on current multipliers, but does not establish dynamic truthfulness and does not repair base-value ranking disagreement.

Figures & tables

	deterministic		calibrated-synthetic
$\Delta$	Even	PI	Even	PI
0	$0.5276\pm 0.0347$	$0.2800\pm 0.0294$	$0.9983\pm 0.0487$	$0.8148\pm 0.0407$
1	$17.7731\pm 0.8702$	$7.4418\pm 0.7871$	$31.8700\pm 1.2762$	$33.1133\pm 1.0317$
2	$50.4252\pm 1.4789$	$47.1974\pm 1.2682$	$72.1369\pm 2.1281$	$70.4868\pm 2.2735$
5	$189.9383\pm 1.8070$	$189.9383\pm 1.8070$	$109.4431\pm 12.0941$	$108.6332\pm 12.6605$
10	$479.2772\pm 2.5343$	$479.2772\pm 2.5343$	$124.0767\pm 4.6931$	$124.0767\pm 4.6931$

Table 1: Pool overspend (%), mean

\pm

two-sided 95% Student-

t

CI half-width. Every cell has

n=30

evaluation seeds 20,260,527–20,260,556,

N=50

devices, 3,000 arrivals/tick, 100 ticks, a mechanism-specific budget frozen from 30 disjoint pilot seeds, and fresh-session frequency-cap state. Even is proportional; PI is proportional-integral.

$\Delta$	empirical (%)	bound (%)	paired slack [95% CI]
1	17.7731	70.4754	52.7024 [51.3435,54.0613]
2	50.4252	145.6363	95.2111 [93.7322,96.6900]
5	189.9383	359.9642	170.0258 [168.2188,171.8328]
10	479.2772	717.1772	237.9001 [235.3658,240.4344]
25	1346.1324	1788.8165	442.6841 [438.9671,446.4011]
50	1669.3075	2357.8528	688.5453 [684.7028,692.3878]

Table 2: Expected-bound evaluation for deterministic Even pacing. Empirical and bound columns are means over the same

n=30

evaluation seeds; “paired slack” is bound minus empirical overspend in percentage points with a two-sided 95% Student-

t

CI. Budgets are frozen from 30 disjoint pilot seeds;

N=50

, 3,000 arrivals/tick, 100 ticks, and fresh sessions in every row.

Appendix figures & tables1 asset

Supplementary material from the paper’s appendix.

Appendix

$\Delta$	original	perturbed
0	$0.5276\pm 0.0347$	$0.5367\pm 0.0414$
1	$17.7731\pm 0.8702$	$18.3370\pm 1.2525$
2	$50.4252\pm 1.4789$	$49.6566\pm 1.0992$
5	$189.9383\pm 1.8070$	$190.3281\pm 1.7013$
10	$479.2772\pm 2.5343$	$480.3457\pm 2.2025$
25	$1346.1324\pm 3.7170$	$1347.0086\pm 3.7100$

Appendix

Table 3: Label-perturbation sensitivity for deterministic Even overspend (%). Entries are mean

\pm

two-sided 95% Student-

t

CI half-width.

Explore similar work

Sep 30, 2026cs.LG

Inference Auctions

When inference demand exceeds available compute capacity, model providers must decide which requests should be served first. Users have different tolerances for delay from an LLM API, but current priority pricing schemes compress these differences into coarse fixed-price service tiers. We design an inference auction that allows users to bid for faster service. Our auction allocates priority in an economically efficient way without sacrificing latency, and we develop fast algorithms for implementing prices that incentivize truthful bidding. We also design an autobidding agent for our inference auction, where users specify an inference budget and the autobidder dynamically adjusts its bids over time to maximize user utility subject to the budget constraint. Experiments validate the practicality of our auction: it increases system welfare while maintaining the cache utilization and latency advantages of SGLang, a state-of-the-art inference serving framework.

Keegan Harris, Siddharth Prasad, Asher Trockman +2

University of California, Berkeley · Toyota Technological Institute at Chicago · Google Research +1

Topic pending

Ask PDF

Apr 29, 2026cs.MA

When Agents Shop for You: Role Coherence in AI-Mediated Markets

Consumers are increasingly delegating purchase decisions to AI agents, providing natural-language descriptions of their preferences and identity. We argue that these representations constitute an information channel, role coherence, through which sellers can infer willingness to pay without explicit disclosure by the buyer agent, leading to preference leakage. In an experiment where a language-model buyer agent shops on behalf of a verbal consumer profile, we show that seller-side inference from dialogue alone recovers willingness to pay nearly one-for-one. Comparing this setting to a numeric-budget condition with confidentiality instructions cleanly isolates role coherence as distinct from instruction-following failure. Because this leakage arises from delegation itself, it cannot be mitigated at the prompt level. Instead, we propose architectural interventions that trade off personalization against preference privacy.

Soogand Alavi, Salar Nozari

Tippie College of Business, University of Iowa

Agentic Commerce Preference Alignment Learning Artificial Intelligence Agents Role Profile

Ask PDF

Jun 28, 2026cs.LG

Learning to Bid in Discriminatory Auctions with Budget Constraints

We study repeated bidding in multi-unit discriminatory (pay-as-bid) auctions for a single bidder with per-round utility equal to value minus

α

times payment, where

α\in[0,1]

is a cost-of-capital parameter. The bidder aims to maximize cumulative utility over

T

rounds subject to a total budget

B

. The problem is challenging even without budgets: the action space is exponential in

M

, the maximum demand of the bidder and the valuation vector (context) varies over time. Exploiting a decomposition of utility across units, we develop polynomial-time learning algorithms based on shortest paths in a directed acyclic graph, obtaining sublinear regret under both full-information and bandit feedback. In the bandit setting, the regret is independent of the number of contexts due to complete cross-learning: observing the utility of the chosen action under the realized context reveals the utility for the same action under all counterfactual contexts. With budget constraints, when the average normalized per-round budget

ρ=\frac{B}{MT}<1

, we design a coupled primal-dual algorithm in which the DAG-based procedure uses dual-adjusted edge weights for primal updates, while online gradient descent updates the dual variable, yielding

ρ

-approximate sublinear regret. Finally, we give implementations whose per-round time and space are independent of the number of contexts, enabling scalability to large or even infinite context spaces.

Negin Golrezaei, Sourav Sahoo

Sloan School of Management, Massachusetts Institute of Technology · Operations Research Center, Massachusetts Institute of Technology

First-Price Auctions Bidding Bandits Token Budget Allocation Vectors

Ask PDF