Beyond monitoring: Envisioning opportunities for AI to help inform biodiversity decision-making
University of Oxford
Friday, June 27, 2025
Biodiversity science and conservation practice have undergone a revolution in seeing, hearing, and sensing the natural world at scale. Many of these advances have been driven by the AI revolution, with machine learning offering increased capabilities to collect and process multi-modal data at scale. However, AI and related computational tools also offer immense capabilities beyond data processing — to determine more effective actions, even under uncertainty, at large scale, and with a complex array of actors and constraints. Such strategic decision-making capabilities of AI have similarly progressed significantly in recent years, but their application has scarcely touched the conservation world despite large potential.
This one-day workshop seeks to identify areas where AI has potential to be leveraged for informing biodiversity actions. We will begin with an overview of the growing potential for algorithmic decision-making methods for biodiversity restoration and conservation, across resource allocation, market design, adaptive management, and impact evaluation. From there, the goal is to identify decisions that are targeted, strategic, and data-driven.
This workshop is sponsored by the Leverhulme Centre for Nature Recovery (LCNR) at Oxford and organized by Lily Xu.
Logistics
The workshop will take place at the School of Geography and the Environment (SoGE) at the University of Oxford. Please check in at the front reception, then follow signs to the Atmosphere room.
Address: S Parks Rd, Oxford OX1 3QY, United Kingdom
Room: Atmosphere room
Getting to Oxford
Schedule
Exact timings subject to change.
Time allocation for talks in (parentheses) before your name.
- 9:30 – 10:00: Coffee and settling in
- 10:00 – 10:20: Welcome! Workshop introduction & quick around-the-room
- 10:20 – 12:20: Session on opportunities in AI. Opportunities in AI for biodiversity decisions, showcase of relevant approaches across AI, economics, and operations research.
(15 min) Tom Dietterich (OSU) - bird migration modeling and uncertainty quantification for invasive species management
(15) Elizabeth Baldwin (Oxford econ) - combinatorial auctions for turtledove habitat
(10) Baizhi Song (LBS) - graph optimization for ocean cleanup
(10) Joseph Tsui (Oxford biology) - disease surveillance at the Kraemer Group
(10) Steven Reece (Oxford LCNR / Smith School) - rainfall run-off with LSTMs, ecoacoustics with embeddings, and ecosystem energetics with entropy theory and graph theory
(10) Samira Barzin (Oxford LCNR / ECI) - geospatial analytics and remote sensing
(8) Lily Xu (Oxford LCNR)
- 12:20 – 13:20: Lunch
- 13:20 – 15:20: Session on opportunities for biodiversity decisions. Showcase from speakers presenting biodiversity decisions or challenges to biodiversity action, followed by discussions about opportunities.
(10 min) Kathy Willis (Oxford LCNR / biology) - AI for policy and business-relevant data
(5) David Benz (Oxford biology) - stacked ecosystem service benefits from public forests
(10) James Askew (NatureVerse) - biodiversity MRV and biodiversity crediting
(12) Felicity Bennett & Paul Turner (Defra) - ELM, data collection, agri-environment monitoring
(10) Prue Addison (BBOWT) - local conservation in Berkshire, Buckinghamshire, and Oxfordshire
(10) Carly Batist (Conservation International) - nature tech applications, unlocking financing, and building AI fluency
(10) Lydia Katsis (ZSL)
(10) Chris Kucharczyk (DrivenData) - Zamba Cloud for camera traps
(10) Alasdair Davies (Arribada) - human-elephant conflict
- 15:20 – 15:40: Coffee break
- 15:40 – 16:15: Group discussions on common themes, key opportunities, shared challenges
- 16:15 – 17:00: Table discussions in structured groups
- 17:00 – 17:10: Quick coffee / stretch
- 17:10 – 17:30: Share takeaways, reflections, closing, next steps
- 17:30 onwards: Informal pub trip @ The King's Arms
40 Holywell Street, Oxford, OX1 3SP
Presentations
Information for presenters:
- Our goal is to help identify for each use case opportunities to leverage AI for more targeted decision-support.
- All presentations are intended to seed discussion, rather than be lectures. I recommend keeping the number of slides small, so that the presentations can be more conversational and adaptive.
Specific for conservation talks:
It would be great to highlight:
- Objectives: Overall goals of the work at your organization or research group
- Intervention types: The types of actions/interventions that are currently being or could be implemented
- Data/infrastructure: Any kind of systematic data collection you use, including specific platforms for data collection or decision support
Hopefully, ideas and insight about "what's possible" will come up throughout the course of the day (which is why the morning is focused on AI opportunities for biodiversity decisions, and then the afternoon is focused on concrete possible use cases). Thus it would help to keep your slides more high-level, to keep it more centered on dialogue rather than a lecture. However, first thinking through your vision for what's possible will help map out where opportunities may lie.
Specific for AI/economics/OR talks:
It would be great to highlight:
- Your field: Broad background to and specific objectives of your technical discipline
- Your research: The specifics of your research and the angle of "interventions" that you explore
- Examples: Any example applications, either from your direct work or others', including from other domains (e.g., healthcare, urban planning)
- Data/infrastructure: Relevant settings, required data/models, or necessary infrastructure to make your work relevant for biodiversity applications
- Any problems you're currently exploring or excited about
Resources
- Hemming et al. (2021). Conservation Biology. An introduction to decision science for conservation
- Possingham et al. (2000). Making Smart Conservation Decisions
- Xu et al. (2023). Reflections from the Workshop on AI-Assisted Decision Making for Conservation
Participants
Alphabetical by organization
- Alasdair Davies (Arribada Initiative)
- Prue Addison (BBOWT)
- Carly Batist (Conservation International)
- Paul Turner (Defra)
- Felicity Bennett (Defra)
- Chris Kucharczyk (DrivenData)
- Millie Chapman (ETH Zürich)
- Baizhi Song (LBS)
- E.J. Milner-Gulland (Oxford LCNR / biology)
- Kathy Willis (Oxford LCNR / biology)
- Lily Xu (Oxford LCNR / economics)
- Samira Barzin (Oxford LCNR / Environmental Change Institute)
- Alex Teytelboym (Oxford LCNR / economics)
- Yadvinder Malhi (Oxford LCNR / geography)
- Steven Reece (Oxford LCNR / Smith School)
- Stephen Thomas (Oxford LCNR)
- Joseph Tsui (Oxford biology & Martin School)
- Seth Flaxman (Oxford CS)
- Abheek Ghosh (Oxford CS)
- Matt Taylor (Oxford CS)
- Elizabeth Baldwin (Oxford economics)
- Paul Klemperer (Oxford economics)
- Chris Summerfield (Oxford experimental psychology)
- Cameron Hepburn (Oxford Smith School)
- Tom Dietterich (Oregon State University)
- James Askew (NatureVerse, Inc.)
- Stephanie O’Donnell (World Bank)
- Ari Lerner (Yale)
- Anthony Dancer (ZSL)
- Lydia Katsis (ZSL)
Contact
Until June 2025: lily.xu@economics.ox.ac.uk
Beyond June 2025: lily.x@columbia.edu
Please note that lily.xu@colum... is NOT me. I do not know who that person is, and any emails sent there will not get to me!
Website designed by Lily. © 2025