AI risk research for decision-makers.

About

I conduct and communicate research to help govern highly capable general-purpose AI. I direct the MIT AI Risk Initiative at MIT FutureTech, and hold positions at MIT FutureTech and The University of Queensland. I'm based in Melbourne.

My methods background is in social psychology, applied behaviour science, implementation science, evidence synthesis, systems thinking, and survey research.

MIT AI Risk Initiative

I direct the AI Risk Initiative at MIT FutureTech. The Initiative is a research project that provides authoritative data and frameworks to help decision-makers across the AI ecosystem identify, prioritise, and manage the most important risks from AI.

• AI Risk Repository

  What are the risks from AI? Which are highest priority?

  A common taxonomy and database synthesising over 1,700 risks from 74 frameworks, used by governments, standards bodies, and AI companies.

• AI Risk Index

  How are leading organisations responding to AI risks?

  Global expert prioritisation of risks and an assessment of how key actors across the ecosystem are managing them.

• AI Mitigations Repository

  What are the mitigations for AI risks? Which are most effective?

  A database of AI risk mitigations extracted from research, policy, and governance frameworks.

• AI Incident Tracking

  Which real-world harms from AI are already occurring?

  A collaboration with the AI Incident Database to monitor and analyse documented incidents of AI harm, and estimate their reach and severity.

• AI Governance Mapping

  Which laws and governance approaches cover which risks?

  A collaboration with the Center for Security and Emerging Technology (CSET) to systematically map governance documents from the AGORA database against risk and mitigation taxonomies.

Australia's AI safety ecosystem

Alongside AIRI, I help build Australia's AI safety ecosystem. I advise Good Ancestors on AI risk and mentor researchers and fellows in the community. I co-founded the Melbourne AI Safety Hub coworking space, and convened the AI Safety Unconferences (2023 and 2025) and the inaugural Australian AI Safety Forum (2024).

Selected publications

Recent

• Saeri, A. K., Graham, J., Noetel, M., Slattery, P., … & Thompson, N. (188 total authors) (2026). Prioritization of Risks from Artificial Intelligence: A Delphi Study of 272 International Experts. arXiv:2606.04490. doi.org/10.48550/arXiv.2606.04490
• Saeri, A. K., Lloyd George, S., Graham, J., Lacarriere, C. D., Slattery, P., Noetel, M., & Thompson, N. (2025). Mapping AI Risk Mitigations: Evidence Scan and Preliminary AI Risk Mitigation Taxonomy. arXiv:2512.11931. doi.org/10.48550/arXiv.2512.11931
• Slattery, P., Saeri, A. K., Grundy, E. A. C., Graham, J., Noetel, M., Uuk, R., Dao, J., Pour, S., Casper, S., & Thompson, N. (2026). The AI Risk Repository: A meta-review, database, and taxonomy of risks from artificial intelligence. Patterns, 7(5), 101517. doi.org/10.1016/j.patter.2026.101517
• Saeri, A. K., Noetel, M., & Graham, J. (2024). Survey Assessing Risks from Artificial Intelligence: Technical Report. UQ / Ready Research. doi.org/10.2139/ssrn.4750953
• Saeri, A. K., & O'Connor, R. M. A. (2023). Applying AI to sustainability policy challenges: A practical playbook. Monash University. doi.org/10.31219/osf.io/y75rq

Earlier

• Saeri, A. K. et al. (2022). What Works to Increase Charitable Donations? A Meta-Review with Meta-Meta-Analysis. VOLUNTAS. doi.org/10.1007/s11266-022-00499-y

Full list: Google Scholar.

Contact

Please contact me to discuss potential collaborations, briefings, or engagements.

Email: [email protected]