30394 research outputs found
Sort by
La donation de l’usufruit de parts de société civile constitue un appauvrissement (Cass. 1ère Civ., 30 avr. 2025, n° 22-20.929)
No full text
Delegation to Artificial Intelligence can increase dishonest behaviour
No full textAlthough artificial intelligence enables productivity gains from delegating tasks to machines1, it may facilitate the delegation of unethical behaviour2. This risk is highly relevant amid the rapid rise of ‘agentic’ artificial intelligence systems3,4. Here we demonstrate this risk by having human principals instruct machine agents to perform tasks with incentives to cheat. Requests for cheating increased when principals could induce machine dishonesty without telling the machine precisely what to do, through supervised learning or high-level goal setting. These effects held whether delegation was voluntary or mandatory. We also examined delegation via natural language to large language models5. Although the cheating requests by principals were not always higher for machine agents than for human agents, compliance diverged sharply: machines were far more likely than human agents to carry out fully unethical instructions. This compliance could be curbed, but usually not eliminated, with the injection of prohibitive, task-specific guardrails. Our results highlight ethical risks in the context of increasingly accessible and powerful machine delegation, and suggest design and policy strategies to mitigate them
Optimal Merger Remedies
Get PDFWe develop a framework to study horizontal mergers when the parties can propose remedies to an antitrust authority. Remedies are modeled as asset divestitures, which make the firm receiving the assets more efficient at the expense of the merged firm. We consider both the case where the merger affects a single market and where it affects multiple markets. Solving for the merging firms’ optimal proposal, we investigate when it involves remedies—and if so, which assets should be divested, and to whom, and how this depends on market characteristics such as the level of competitiveness
Heterogeneous preferences and asymmetric insights for AI use among welfare claimants and non-claimants
Get PDFThe deployment of AI in welfare benefit allocation accelerates decision-making but has led to unfair denials and false fraud accusations. In the US and UK (N = 3,249), we examine public acceptability of speed-accuracy trade-offs among claimants and non-claimants. While the public generally tolerates modest accuracy losses for faster decisions, claimants are less willing to accept AI in welfare systems, raising concerns that using aggregate data for calibration could misalign policies with the preferences of those most affected. Our study further uncovers asymmetric insights between claimants and non-claimants. Non-claimants overestimate claimants’ willingness to accept speed-accuracy trade-offs, even when financially incentivized for accurate perspective-taking. This suggests that policy decisions aimed at supporting vulnerable groups may need to incorporate minority voices beyond popular opinion, as non-claimants may not easily understand claimants’ perspectives. This work highlights the importance of stakeholder engagement and transparent communication in government deployment of AI, particularly in power-imbalanced contexts
The science fiction science method
No full textPredicting the social and behavioural impact of future technologies before they are achieved would enable us to guide their development and regulation before these impacts get entrenched. Traditionally, this prediction has relied on qualitative, narrative methods. Here we describe a method that uses experimental methods to simulate future technologies and collect quantitative measures of the attitudes and behaviours of participants assigned to controlled variations of the future. We call this method ‘science fiction science’. We suggest that the reason that this method has not been fully embraced yet, despite its potential benefits, is that experimental scientists may be reluctant to engage in work that faces such serious validity threats. To address these threats, we consider possible constraints on the types of technology that science fiction science may study, as well as the unconventional, immersive methods that it may require. We seek to provide perspective on the reasons why this method has been marginalized for so long, the benefits it would bring if it could be built on strong yet unusual methods, and how we can normalize these methods to help the diverse community of science fiction scientists to engage in a virtuous cycle of validity improvement
Gaussian Agency problems with memory and Linear Contracts
No full textCan a principal still offer optimal dynamic contracts that are linear in end-of-period outcomes when the agent controls a process that exhibits memory? We provide a positive answer by considering a general Gaussian setting where the output dynamics are not necessarily semimartingales or Markov processes. We introduce a rich class of principal–agent models that encompasses dynamic agency models with memory. From a mathematical point of view, we show how contracting problems with Gaussian Volterra outcomes can be transformed into those of semimartingale outcomes by some change of variables to use the martingale optimality principle. Our main contribution is to show that for one-dimensional models, this setting always allows optimal linear contracts in end-of-period observable outcomes with a deterministic optimal level of effort. In higher dimensions, we show that linear contracts are still optimal when the effort cost function is radial, and we quantify the gap between linear contracts and optimal contracts for more general quadratic costs of efforts
Information Disclosure in Preemption Races: Blessing or (Winner's) Curse?
No full textTwo players receiving independent signals on a risky project with common value compete to be the first to innovate. We characterize the equilibrium of this preemption game as the publicity of signals varies. Private signals create a winner's curse: investing first implies that the rival has abstained from investing, possibly because he has privately received adverse information about the project. Since players want to gather more evidence in support of the project as a compensation, they invest later when signals are more likely to be private. Because of preemption, the NPV of investment is zero at equilibrium regardless of the publicity of signals. However, for a conservative planner who cares about avoiding unprotable investments, this implies that investment arises too early at equilibrium, and such a planner then prefers signals to be private. This provides a rationale against the mandatory disclosure of negative results in science, notably when competition is severe. Our results suggest that policy interventions should primarily tackle winner-takes-all competition, and regulate transparency only once competition is suciently mild
Childhood skeletal lesions common in prehistory are present in living forager-farmers and predict adult markers of immune function
No full textPorous cranial lesions (cribra cranii and cribra orbitalia) are widely used by archaeologists as skeletal markers of poor child health. However, their use has not been validated with systematic data from contemporary populations, where there has been little evidence of these lesions or their health relevance. Using 375 in vivo computed tomography scans from a cohort-representative sample of adults aged 40+ years from the Bolivian Amazon, among food-limited, high-mortality forager-farmers, we identified cribra cranii on 46 (12.3%) and cribra orbitalia on 23 (6%). Cribra orbitalia was associated with several hallmarks of compromised immune function, including fewer B cells, fewer naïve CD4+ T cells, a lower CD4+/CD8+ T cell ratio, and higher tuberculosis risk. However, neither lesion type predicted other physician-diagnosed respiratory diseases, other markers of cell-mediated immunity, or hemoglobin values. While cribra orbitalia shows promise as a skeletal indicator of health challenges, our findings do not support the continued practice of using these lesions to infer anemia in adults