Professor working at the intersection of computational social science, behavioural science, and public policy.
What I study
Designing and analysing randomised controlled trials for behavioural interventions and public policy. Translating evidence into deliverable strategies through frameworks for time-sensitive interventions, personalised interventions, and implementation science.
Applying computational methods—from sequence analysis, network models, and agent-based models to large language models and AI—to understand social phenomena and human behaviour at scale.
Experimental design, online experiments, multiverse analysis, metascience, and quantitative methods for social science. Author of open-source R packages including multiverseRCT, BriDGE, and sequenceRCT.
Journal of Artificial Societies and Social Simulation, 29(3), 1.
When structural assumptions hold, targeted system-level seeding is markedly more transformative than individual-level nudges—yet the same reliance on social reinforcement makes it fragile.
Debate over whether behavioural policy should target individual-level (i-frame) or system-level (s-frame) interventions has been rich but largely qualitative. This article addresses that gap with an agent-based model, comparing i-frame interventions—nudges, information campaigns, rotating micro-targeting—with s-frame interventions—uniform structural levers and targeted seeding—across three sources of misspecification: imperfect structural knowledge, heterogeneous response, and external shocks. Two findings stand out. When structural assumptions are approximately correct, targeted s-frame seeding can trigger self-reinforcing diffusion and is markedly more transformative and cost-efficient than i-frame benchmarks. Yet that same reliance on social reinforcement creates fragility: targeting errors and backsliding shocks reduce s-frame performance more sharply. The contribution is a mechanism-oriented framework for comparing intervention families under uncertainty and for making the transformation–robustness trade-off explicit in behavioural policy design.
Read ArticleSelected work
Veltri, G. A., & Acerbi, A. (2026). Social Learning Rules and the Effectiveness of Behavioural Policy: An Agent-Based Model. Behavioural Public Policy. In press.
Veltri, G. A., Mureddu, F., Innocenti, A., Sirizzotti, M., & Venturini, E. (2026). Regulatory frameworks and digital identity in the metaverse: an exploratory virtual reality experiment on trust, privacy, and openness. Frontiers in Virtual Reality, 7:1779260. DOI
Veltri, G. A. (2026). S-Frame vs. I-Frame Interventions: The Transformation-Robustness Trade-Off in Behavioral Policy. Journal of Artificial Societies and Social Simulation, 29(3), 1. DOI
Veltri, G. A. (2026). Navigating evidence, legitimacy and delivery: A three-dimensional framework for behavioural policy design. Policy and Society. DOI
Veltri, G.A.(2026). Time-Sensitive RCTs in Behavioural Public Policy: A Pragmatic Framework Using Sequence Methods, Personalisation, and Reinforcement Learning. Frontiers in Behavioral Economics. In press. DOI
De Duro, E. S., Franchino, E., Improta, R., Veltri, G. A., & Stella, M. (2026). Cognitive networks identify AI biases on societal issues in Large Language Models. EPJ Data Science. DOI
Veltri, G.A. (2025). The Effects of Data Preprocessing Choices on Behavioral RCT Outcomes: A Multiverse Analysis. Multivariate Behavioral Research. DOI
Veltri, G.A. (2025). From Evidence to Delivery: An Implementation‑Science Blueprint for Behavioural Policy. Behavioural Public Policy. DOI
Andrei, F., & Veltri, G. A. (2025). Signalling strategies and opportunistic behaviour: Insights from dark-net markets. PLOS ONE, 20(3), e0319794. DOI
Work in progress
Carrillo, A., Citraro, S., Ardebili, A. A., Taietta, E., Rossetti, G., Ferrara, E., Veltri, G. A., & Stella, M. (2026). LLMs can persuade only psychologically susceptible humans on societal issues, via trust in AI and emotional appeals, amid logical fallacies. arXiv. Preprint
Veltri, G. A., & Gilbert, J. (2026). Results from Randomized Controlled Trials are Highly Sensitive to Data Preprocessing Decisions: A Multiverse Analysis of 97 Outcomes. MetaArXiv. Preprint
Carrillo, A., Taietta, E., Ardebili, A. A., Veltri, G. A., & Stella, M. (2026). Talk2AI: A Longitudinal Dataset of Human–AI Persuasive Conversations. arXiv. Preprint
Andrei, F., Tizzoni, M., & Veltri, G. A. (2026). Dengue risk perception and public preferences for vector control in Italy and France: utility and regret-based choice experiments. medRxiv. Preprint