Transitioning from FAIR to AI Ready Data in the Physical Sciences

In recent years, the physical sciences community has been generating increasingly large and complex datasets, at a scale that is now beyond what can be fully explored or analysed by humans alone. As a result, researchers are turning to AI and machine‑learning techniques, which have matured significantly and offer powerful new ways to extract insight from data. However, while the adoption of FAIR data principles has improved data sharing and reuse, experience is showing that FAIR does not necessarily mean AI‑ready. Many datasets remain difficult to use effectively in AI and Machine Learning models.

This interactive workshop has been co-created by the Physical Sciences Data Infrastructure (PSDI) and the AI in Chemistry Hub (AIchemy). It aims to bring together researchers, data professional and infrastructure developers to facilitate knowledge exchange and explore what it truly means to be “AI Ready”. The workshop is comprised of invited presentations, lightning talks from participants and interactive discussion sessions. The talks will share current practices, highlighting successes and challenges, and the discussion sessions will explore the practical approaches and tools for evaluating and improving AI readiness.

Audience

This in-person event is aimed at anyone interested in dataset standards, curation, and developing robust methods to assess the applicability and reliability of data for reuse. It will be particularly relevant for researchers and research software engineers working with data and AI/ML, data stewards and research data managers, infrastructure and platform developers, and scientists interested in enabling future reuse of their datasets.

Call for Lightning Talks

We invite submissions for short lightning talks exploring the challenges, opportunities, and practical experiences involved in creating AI-ready datasets within the physical sciences.

This is an opportunity to share emerging ideas, real-world case studies, and lessons learned from working with data intended for AI and machine-learning applications.

Topics may include, but are not limited to:

• Experiences in developing or curating AI-ready datasets
• Challenges in preparing FAIR data for AI and machine learning use
• Data quality, metadata, interoperability, and standardisation
• Benchmarking, validation, and reproducibility
• Infrastructure, tooling, and workflow development
• Community needs, open challenges, and proposed solutions

During registration, participants will be able to indicate their interest in presenting a lightning talk. The organising team will review submissions and notify successful applicants by 26 June 2026.

Draft Agenda

• 10:00-10:30: Coffee & Registration

• 10:30-10:35: Housekeeping & Welcome

• 10:35-10:45: Introduction to PSDI

• 10:45-10:55: Introduction to AIchemy

• 10:55-11:10: Setting the Scene: From FAIR to AI-Ready

• 11:10-11:25: Coffee Break & Networking

• 11:25-12:45: Invited Speakers (Matthew Partridge, Aileen Day, Nessa Carson)

• 12:45-13:30: Networking Lunch

• 13:30-14:00: Participant Lightning Talks

• 14:00-14:15: Introduction to Discussion Sessions

• 14:15-14:45: Discussion Sessions Part 1

• 14:45-15:00: Coffee Break & Networking

• 15:00-15:30: Discussion Sessions Part 2

• 15:30-16:00: Feedback & Wrap Up

Event Travel

The University of Southampton is accessible via various different transport links

• Travelling by Train: Southampton Airport Parkway is the closest station, but Highfield Campus is also close to St Denys and Southampton Central Station

• Travelling by Bus: Highfield Campus is on the bus route for all Unilink Busses

• Travelling by Car: Highfield Campus has very limited parking, if you require on-site parking for accessibility reasons then please note that in your registration, or email [email protected] at least one week in advance, and we will try and accommodate this if possible (NB: we cannot guarantee this in the week leading up to the event). There is also a 100-space short stay carpark (with charges), but this is likely to be very busy.

Know what’s Happening Next — before everyone else does.

Get App

Host or Publisher Physical Sciences Data Infrastructure (PSDI)