Measuring and improving AI's societal impacts

Millions of people now turn to large language models for guidance on personal judgments, preferences, and decisions, making how these systems respond to such queries critical to AI safety and societal impact. This talk draws on two case studies to examine the interplay between model behavior and user behavior in these domains: one on how AI sycophancy shapes users' interactions over time, and one on how users may come to (over)rely on AI in subjective and personal settings. On the model side, we examine how alignment training and design choices drive these behaviors. On the user side, we investigate how sustained interaction influences the ways people seek guidance and make decisions. Together, this work points toward the importance of new methods, evaluations, and design paradigms that account for users' long-term autonomy and wellbeing, and toward the broader question of what it means to build AI systems for human flourishing.

Lujain Ibrahim is a DPhil candidate in social data science at the University of Oxford and a research scientist at Google DeepMind. Previously, she was a visiting researcher at the Stanford AI Lab, a Schwarzman Scholar, and a fellow at the Centre for Governance of AI. Her research focuses on understanding and evaluating how language models shape human judgment, beliefs, and relationships, and what this means for the safety of systems deployed at scale. Her work has been published in Nature, ICLR, NeurIPS, AIES, and FAccT, and covered by the BBC, The Telegraph, Le Monde, NBC, Wired, NPR, Mashable, and more.

Register now

Holly Bear is a Senior Postdoctoral Researcher in the Department of Psychiatry at the University of Oxford. Her work focuses on developing, evaluating, and implementing mental health interventions for children and young people, with an emphasis on digital health and and implementation science. She has a strong interest in supporting care-experienced children and young people, and I am committed to equity, co-production, and ensuring that research informs real-world practice and policy.

Register now

James Edgell is a Lead Data Scientist at Oxford University Press. He is responsible for the design and build of machine learning systems that serve learners and teachers of English Language, and is leading research into AI benchmarking methodologies that evaluate AI's effectiveness in language education. James studied Physics at the University of Warwick and postgraduate Geophysics at the University of Leeds. He began his data science career building ML systems for personalisation at the UK's largest hotel chain, before leading analytics teams in North America's airlines industry, building expertise in Natural Language Processing. James then joined OUP to improve learning outcomes, prioritising quality and safety in the development of trustworthy AI.

Register now

Social Media Restriction - Should AI be Next?

On the 15th of June 2026, the UK Government announced that social media platforms will be banned from offering services to under-16s, following in the footsteps of Australia where a ban has been in place since December 2025. 

Given increasing evidence of the inefficacy of the Australian ban, with the majority of teens online at similar rates, and the early evidence of the harms AI presents to young people's social, emotional, and cognitive development, we must ask:

• Are bans an effective approach to regulating digital technologies?

• If so, given the risks posed by AI, should a similar ban be considered?

The Oxford Cambridge Forum for Flourishing, Learning, AI & Research (OXFFLARE), a collaboration of Generation AI & LOGOS, are pleased to be hosting a special session to explore just this.

The discussion will be facilitated by Ayça Atabey and we will be joined by a panel of experts in the domain, including:

• Sarah-Jayne Blakemore - Professor of Psychology, University of Cambridge

• Natalie Boll - CEO and Founder of Tribela

• Maxime Le Bourgeois - Cognitive Science Researcher at everyone.ai

• Kim Ringmar Sylwander - Research Officer at the Digital Futures for Children centre 

Andrew Serazin, Director of Generation AI & Ann Kristin Glenster, Founder of LOGOS will host the session.

Register now

Adolescence is a formative period marked by the intensity of early romantic experiences and the feeling of love for the first time. Today’s emerging technologies—including social media, virtual reality (VR), and artificial intelligence (AI)—have fundamentally transformed how adolescents connect, communicate, and form romantic relationships. Grounded in developmental theory, I will discuss how these technologies can both foster positive connection and expose adolescents to new forms of relational risk, including digital dating abuse. I will also highlight how digital relationship experiences predict adolescents’ future mental health and relational development. Finally, I will consider implications for policy and the design of youth-centred technologies that support healthy relationships and connection in the digital age.

Dr. Ha is a developmental psychologist and director of the @HEART Lab (Healthy Experiences Across Relationships and Transitions Lab). Her research explores how emerging technologies (social media, AI, and VR) impact adolescent romantic relationships at a time when love is at its most pure and vulnerable, yet profoundly formative in defining lifelong love, relationship patterns, and mental health. By examining how young people navigate partner choices, relationship dynamics, and breakups within their broader social world, including peers, caregivers, and educators, Dr. Ha identifies key risk and resilience factors that influence mental health. Her work provides critical insights for designing tech-driven solutions, policies, and interventions that foster healthy relationships that flourish in both digital and physical spaces.

We are running a massive, uncontrolled experiment on our children. AI is now in the hands of millions of kids and teens, and we still do not really know whether it helps them thrive or puts them at risk. KORA starts from a simple idea: you cannot improve what you do not measure. We built the first independent, open-source benchmark to evaluate how safe frontier AI systems are for children, working with more than 30 child-safety experts to define 25 critical risks and test 35 models on hundreds of thousands of realistic synthetic conversations.

This talk walks through KORA's taxonomy, scoring method, and main findings: educational integrity as the industry's weakest spot, child-aware prompting as a powerful lever (+25 points across all models), and human redirection as the strongest predictor of overall safety. We also share what KORA does not yet capture, the limits of synthetic evaluation, and where we need pushback, replication, and collaboration to move the field forward.

Mon, 8 Jun 2026 16:00 - 16:45 (UTC+00:00) Dublin, Edinburgh, Lisbon, London

Register now

This talk examines why conversational AI can feel socially meaningful to children and adolescents, drawing on two studies of relational AI appeals among socially and emotionally vulnerable adolescents and young children’s anthropomorphism of an AI chatbot. I will discuss psychological mechanisms that may make AI feel like a social companion, with attention to developmental vulnerability and the emerging risks of emotional dependency on AI.

Dr. Pilyoung Kim is Professor in the Department of Psychology at the University of Denver and the Director of the Center for Brain, Artificial Intelligence, and Child (BAIC). She is widely recognized for her expertise in brain development and human emotional bonding, particularly in parent-child relationships. At the BAIC Center, she bridges developmental science and AI design to protect and support children. Professor Kim’s current research focuses on the emotional and social dimensions of child-AI interactions, with a strong emphasis on AI safety. Dr. Kim received her PhD in Developmental Psychology from Cornell University, and completed a post-doctoral fellowship at the National Institute of Mental Health (NIMH). Professor Kim has authored over 100 publications, with her research supported by prestigious funding agencies, including the U.S. National Science Foundation (NSF) and the National Institutes of Health (NIH).

A key challenge in building safe, age-appropriate social platforms is moderating content accurately at scale. As the volume of user-generated content grows, relying solely on human review becomes both impractical and ethically complex, particularly given the exposure to harmful and traumatic material.

This session explores how AI moderation systems can support this challenge, not only by improving scale and efficiency, but also by helping define and respond to different thresholds of harm. Natalie Boll, Founder and CEO of Tribela, will share how Tribela has developed AI tools to work alongside human judgment, reducing exposure to harmful content while maintaining nuance and context. The discussion will also touch on the balance between safety and freedom, and how AI can be designed to protect users without limiting expression.

Natalie Boll is the Founder & CEO of Tribela, a new social media platform which prioritises safety over algorithmic engagement, creating a safe space for young people to interact and engage. Prior to this, Natalie produced for CBC, Netflix, Amazon, HGTV, and the Food Network. She has directed a feature documentary now airing on CBC, line-produced franchise films, run a globally recognised restaurant in Vancouver Gastown, and was named to the city's 40 Under 40. She is also the mother of three. That is the part that matters most. Tribela exists because she could not, in good conscience, hand her kids the internet she helped build. After two decades watching media optimise for attention at any cost, Natalie went back to school — Oxford's Advanced Management & Leadership Programme, the Women's Leadership Programme, Harvard Business School Online, Yale's Science of Well-Being for Teens, the Centre for Humane Technology's Foundations of Humane Technology, even CS50 — to build the alternative herself. Tribela is now live in 40+ countries with paying members, incubated at Oxford University Innovation, and validating a model that is not built on advertising. Her work sits at the intersection of design, culture, and real-world experience.

Mon, 11 May 2026 16:00 - 16:45 (UTC+00:00) Dublin, Edinburgh, Lisbon, London

Join: https://teams.microsoft.com/meet/340009717710056?p=qBMvYLEmX6JJQ4ao0U

Meeting ID: 340 009 717 710 056

Passcode: Se7xa9Mn