9 September 2024 · Imperial College London
Human & Artificial
Intelligence
in Organization
Interdisciplinary Impact
13:00–17:30 · I-X Hub
White City, London
The lineup
Speakers
and
respondents
Organizing
committee
Monday · 9 September 2024
Schedule
13:00–17:30
All times London time (BST)
- –
Opening remarks
Kevin Corley (Imperial)
- –
Session 1 · Unraveling the Synergies and Differences Between Human and Artificial Intelligence
Presenters: Kevin R. McKee (Google DeepMind), Winnie Street (Google Research). Respondent: Martin Anthony (LSE).
- –
Coffee break
- –
Session 2 · Building Trustworthy AI: Balancing Openness, Ethics, Reliability, and Transparency
Presenters: Jennifer Ding (The Alan Turing Institute), Lujain Ibrahim (Oxford). Respondent: Aidan Peppin (Cohere for AI).
- –
Panel discussion
Moderator: Susan Scott (Imperial)
- –
Closing remarks
David Shrier (Trusted AI Alliance)
- –
Post-event reception
Session 1
Unraveling the Synergies and Differences Between Human and Artificial Intelligence
Paper presentation
Explore how social perception influences human-AI cooperation. This study reveals that warmth and competence assessments of AI agents better predict human preferences than task performance, emphasizing the importance of social factors in human-AI interactions.
Paper presentation
Assessing LLM performance on higher-order theory of mind tasks
Winnie Street (Google Research)
Read paperDiscover how the performance of large language models (LLMs) compares to that of humans on theory of mind (ToM) tasks which involve reasoning about the multiple interrelated mental states of characters in short stories. This study shows that larger, fine-tuned LLMs come close to—and sometimes exceed—human performance on these tasks, and explores the implications for user-facing LLM applications and human-AI interactions.
Respondent-led discussion
Synergies and Differences Between Human and Artificial Intelligence
Respondent: Martin Anthony (LSE)
Explore the unique strengths and complementary roles of human and artificial intelligence. This panel examines recent findings on AI's theory of mind capabilities and discusses how AI can augment human cognition across various domains.
Session 2
Building Trustworthy AI: Balancing Openness, Ethics, Reliability, and Transparency
Paper presentation
Expanding Participatory AI Collaboration & Governance
Jennifer Ding (The Alan Turing Institute)
Read paperUncover collaboration patterns in the open AI ecosystem through a quantitative analysis of the Hugging Face Hub. Learn how licenses impact development activity and the role of key actors in shaping the future of AI.
Paper presentation
Beyond static AI evaluations: advancing human interaction evaluations for LLM harms and risks
Lujain Ibrahim (Oxford)
Read paperExplore the need for human interaction evaluations in assessing AI safety. Learn about a new framework for evaluating human-model interactions, crucial for understanding and mitigating potential harms in real-world AI applications.
Respondent-led discussion
Fostering Openness, Transparency, and Ethical Standards in AI Development
Respondent: Aidan Peppin (Cohere for AI)
What are strategies for fostering openness, transparency, and ethical standards in AI development? This panel explores collaborative practices in the open AI ecosystem and discusses the importance of multi-stakeholder collaboration in shaping future AI policies.
White City · London
Getting there
I-X Hub
6th floor, Translation & Innovation Hub (I-Hub)
84 Wood Lane, London W12 0BZ
Upon arrival, check in at the reception desk. The staff will provide you with a visitor pass and direct you to the correct floor for the symposium.
By car
Just off the A40 (Westway). Taxis and Ubers can drop you off directly at the entrance of the I-Hub.
By bus
White City Bus Station: roughly a 12-minute walk.
Cavell House Bus Stop (Stop P): about a 5-minute walk.
By Tube
White City Station (Central Line): about a 5-minute walk.
Wood Lane Station (Hammersmith & City and Circle Lines): roughly a 10-minute walk.