Past Events
Talking to Machines Seminar Series
Friday, 13th February 2026
14.00 – 15.00, hybrid format
Butler Room, Nuffield College & Zoom
Brian Scholl
Dr. Brian Scholl is a leading economist at the intersection of regulation, financial markets, and artificial intelligence. He currently serves as a Staff Regulatory Researcher at Norm Ai, where he leads work on quantifying risk and designing data-driven, AI-enabled compliance programs for financial institutions.
Previously, Dr. Scholl was the founding Chief Economist of the U.S. Securities and Exchange Commission Office of Investor Research, where he led cutting-edge research on investor behavior and emerging technologies, including AI. He also served as Chief Economist of the U.S. Senate Budget Committee, advising policymakers on economic evidence and fiscal policy.
Recognized as the U.S. government’s top evidence innovator in 2022, Dr. Scholl pioneered rapid-cycle evidence generation systems that dramatically shortened research timelines and translated economic insights into real-world policy impact. Across roles, he has been known for building bridges between regulators, markets, and technology—turning complex economic and behavioral research into actionable frameworks.
Title: Experimental Evidence on Decision-Making: Implications for Governance in an AI-Enabled Future
Abstract
Rapid advances in artificial intelligence have renewed longstanding questions about human decision-making, institutional design, and governance. Is this technological moment fundamentally different? How do new systems interact with human cognitive limitations? And under what conditions do they enhance—or undermine—trust, legitimacy, and social welfare?
This talk draws on a series of large-scale behavioural experiments I conducted in the context of financial regulation to examine how individuals make consequential decisions in environments characterized by complexity, asymmetric information, and institutional mediation. Retail investors routinely face choices—whether to participate in markets, how to allocate assets, how to interpret disclosures, and whether to rely on advice—that exceed the capacities of unaided human cognition. Traditional regulatory tools, particularly disclosure, have struggled to address these challenges and, in important respects, have fallen short of their intended protective role.
I review evidence from multiple randomized experiments that investigate how features of choice architecture shape behaviour, including linguistic complexity and jargon, reference points and performance benchmarks, and reliance on expert advice. Across settings, the findings reveal both the promise and the fragility of behavioural interventions: modest changes in presentation can meaningfully improve decisions, yet the same mechanisms can also be exploited by intermediaries facing competing incentives. In one set of studies, simplifying language and carefully structuring choice environments substantially improves comprehension and decision quality; in others, benchmark performance framing strongly influences investment choices even when underlying fundamentals are unchanged. A further experiment shows that individuals exhibit limited ability to screen the quality of financial advice, accepting poor guidance nearly as often as good guidance.
Taken together, these results highlight persistent limits to individual self-protection in complex systems and raise broader questions for the governance of AI-enabled decision support. While AI systems hold the potential to augment human capital, triage information, and personalize guidance at scale—benefiting individuals, institutions, and regulators alike—they also risk amplifying manipulation, opacity, and power asymmetries if left unchecked.
The talk concludes by using these experimental findings as a lens to prompt discussion about the design and regulation of AI-enabled institutions. Rather than treating AI as a replacement for human judgment, the discussion emphasizes how evidence on human behaviour can inform regulatory and institutional frameworks that are more trustworthy, legitimate, and aligned with long-run systemic stability—while acknowledging the risks such systems pose and the possibility that fundamentally new governance approaches may be required.
Talking to Machines Seminar Series
Friday, 21st November 2025
15.00 – 16.00, hybrid format
Lecture Room, Nuffield College & Zoom
Mohsen Mosleh
Mohsen is an Associate Professor at Oxford Internet Institute, a Governing Body Fellow at Wolfson College, and a research affiliate at MIT Sloan School of Management. His lies at the intersection of computational (data science) and cognitive psychology. He studies how information spreads on social media and how ties are formed on social networks.
Title: Echo Platforms & Conversational Corrections
Abstract
In this talk, I will talk about two recent projects.
First, I will talk about a cross‑platform study where we collected 10 million news‑link posts across seven social media platforms and examined engagement by outlets’ political slant and quality. We find that platforms with more conservative user bases share lower‑quality news on average; the partisan engagement advantage flips by platform (aligned content wins); and within users, lower‑quality links get higher average engagement — even on non‑ranked feeds — implicating user preferences rather than algorithms.
Next, I will talk about an experimental project that combines lab and field to debunk false claims using LLMs. We tune conversational corrections in a simulated feed, then deploy the best strategies on Bluesky and X to test external validity with native engagement metrics.
Wednesday, 25th June
Carlos III University – Madrid, Spain
Livia Schubiger (D-GESS, ETH Zurich) & Raymond Duch (University of Oxford) organized an EPSA Madrid 2025 pre-conference that assembled researchers working with Large Language Models (LLMs). The focus were on the application of LLMs to research being conducted in the social sciences.
Delivering Health Innovation in Resource-Constrained Settings: Medical Breakthroughs into Development Outcomes
From cutting-edge vaccine trials to behavioural insights for public health and global policy debates — featuring leading researchers from the Department of Economics, University of Oxford, Blavatnik School of Government, University of Oxford, King’s College London, and Gavi, the Vaccine Alliance.
- Date: Monday, 23rd June 2025, from 1pm.
- Location: Manor Road Building, Oxford, OX1 3UQ.
- Speakers: Niccolò Francesco Meriggi, Mehreen Datoo, Katy Kuhrt, Alice Beardmore-Gray, Ray Duch, Stefan Dercon, David Kinder CBE, Dan Hogan
- More info: https://economics.web.ox.ac.uk/research-jamboree-2025
Synthetic Replication Games Workshop
The IMEBESS 2025 Synthetic Replication Games was a one-day in-person workshop organised by the Talking to Machines team in the context of a large replication project exploring the potential of large language models (LLMs) to augment human samples in experimental social science. This workshop invited researchers of all backgrounds and career stages to collaborate in adapting experimental studies from top-tier journals into a standardised format and replicate them using a variety of LLMs. The results were presented in a dedicated round table at IMEBESS 2025, and contributors were recognised as co-authors on the final publication.
The workshop was held on 21th May 2025 in the IMEBESS conference venue.
Workshop on the 2024 U.S. Presential Election Polling
- Date: Monday, 25th November 2024
- Time: 9.30am – 5pm
- Location: Nuffield College, University of Oxford, Oxford (UK)
The post-election workshop was an opportunity to critically assess the various data collection and modeling strategies employed for pre-election estimates of the vote share of the major U.S. presidential candidates over the course of the campaign. Leading pollsters from academia and industry critically assessed the state of pre-election polling in the aftermath of the 2024 Presidential Elections.
| 9:00am – 9:30am | Coffee |
| 9:30am – 10:00am | Welcome and Introductions |
| 10:00am – 11:00am | National Election Studies & Commercial Polling – Shanto Iyengar, Professor (Stanford) and Principal Investigator of the American National Election Study (ANES) – Jane Green, Professor (Nuffield College), President of the British Polling Council, and Principal Investigator British Election Study – Clifford Young, President Public Affairs (IPSOS U.S.A.) |
| 11:00am – 11:30am | Coffee Break |
| 11:30am – 12:30pm | Advances in Model-Based Election Election-Polling – Lucas Leemann, Professor (University of Zurich) – Kosuke Imai, Professor (Harvard University) |
| 12:30pm – 1:30pm | Lunch |
| 1:30pm – 2:30pm | Polling Aggregation: Resilience to Systematic Bias – Eli McKown-Dawson, LSE and Silver Bulletin – Steve Fisher, Professor (University of Oxford) – Martin Strabe and Oliver Hawkins (The Financial Times) |
| 2:30pm – 2:45pm | Coffee |
| 2:45pm – 3:45pm | Election Forecasting: Mapping the State of the Art – Mary Steigmeier, Professor (University of Missouri) – Philippe Mongrain, Post-doctoral Researcher (University of Antwerp) |
| 3:45pm – 4:00pm | Coffee |
| 4:00pm – 5:00pm | Silicon Sampling for Public Opinion Polling – Roberto Cerina, Assistant Professor (University of Amsterdam) – Ray Duch, Professor (Nuffield College) – Ben Warner, Founder (Electric Twin) |