An independent AI safety policy think tank

Risk research to guide
AI safety decisions.

Acadelica is an independent policy think tank focused on AI safety and policy. Through research into innovation risk, we aim to inform policies that help AI labs and governments reduce catastrophic AI risk. We are also exploring how law firms can use AI safely and contribute legal expertise to wider AI safety efforts.

Explore our purpose

Risk research to
inform AI safety policy.

Acadelica was founded by Ariana Strait to advance research into AI risk and policy, with the aim of minimising the likelihood of catastrophic harm.

Policy

Acadelica studies how AI innovation affects risk and uses its findings to develop policy guidance that supports AI safety.

Education

We make research on AI risks accessible to the public, helping people understand what AI developments could mean for society and how policy could address the risks they pose.

Our approach

Research into AI risk to inform policy
and reduce the likelihood of catastrophic harm.

Understanding how
AI innovation creates risk

Risk analysis examines how innovation affects a system and why emerging risks may be difficult to recognise or contain. MIRC provides a framework for tracing how innovation can create, disguise and transmit risk.

Applying this analysis to AI means examining how advances in capabilities and decisions about their use could lead to harm. The aim is to identify where safeguards may be needed.

Developing policy
from risk analysis

Risk analysis helps identify where policy could prevent harm. For AI labs, this means examining development and deployment decisions. For governments, it means considering where public oversight is needed.

Policy proposals should explain how a recommended measure addresses the identified risk, drawing on the evidence and acknowledging uncertainty.

Research and Publications

Risk research

Applying MIRC to AI Risk and Policy

Applying the Matrix of Innovation-Risk-Crisis to AI innovation as a basis for tailored safety policy.

Recursive Self-Improvement and Catastrophic AI Risk

Applying MIRC to examine how recursive self-improvement could contribute to catastrophic harm.

All views expressed are personal and independent of any organisational affiliation.

AI Labs

Policy research for AI labs examines how decisions about developing and deploying advanced AI affect risk. It aims to identify policies that could reduce the likelihood of catastrophic harm.

Governments

Policy research for governments examines how public oversight could reduce catastrophic AI risk. It focuses on risks that individual organisations cannot address alone and where government intervention could help.

Law Firms and the Legal Industry

Acadelica explores how law firms and the wider legal industry can contribute to AI safety through legal expertise, pro bono support, policy research and partnerships with organisations working to reduce AI risk.