AI Awards
Monthly shortlists and annual recognition of excellence in artificial intelligence.
AICI Responsible AI Award 2025
The second annual AICI Responsible AI Award. As in 2024, AICI awards organisations that demonstrate responsible AI practice through evidence rather than assertion. The 2025 edition pays particular attention to the EU AI Act's enforcement period: organisations that have invested in genuine conformity — not box-ticking — are eligible. Organisations that have used the compliance process as a rebranding exercise are not.
AICI Policy Leadership Award 2025
The second annual AICI Policy Leadership Award. The 2025 edition recognises contributions to AI governance at a particularly consequential moment: the EU AI Act entering enforcement, the Council of Europe Framework Convention open for ratification, and the first generation of national AI laws taking effect across the OECD. AICI is looking for evidence of durable governance outcomes, not conference presentations.
AICI Responsible AI Award 2024
Awarded to the organisation that has most demonstrably embedded responsible AI practice into its operations — not through policy statements, but through evidence of changed decision-making, meaningful oversight mechanisms, and accountability when things go wrong. AICI looks for practice over posture. Nominations are evaluated on documentation of governance processes, evidence of human oversight in high-stakes AI decisions, and the organisation's response to AI incidents or failures. An organisation that handled a failure well is more eligible than one that has not yet failed.
AICI AI Safety Research Award 2024
Awarded to the individual researcher or research group whose published work in 2024 made the most significant contribution to AI safety — defined as reducing the probability of AI systems causing unintended harm, whether through misalignment, misuse, or unforeseen failure modes. The award recognises technical safety research: interpretability, robustness, alignment, evaluation methodology. Governance and policy work is recognised under the Policy Leadership Award.
AICI Policy Leadership Award 2024
Awarded to the individual who has most effectively translated AI research into policy outcomes — legislation drafted or improved, regulatory frameworks strengthened, international agreements advanced. AICI recognises that the gap between AI research and AI governance is a structural problem: the people who understand AI well rarely have policy influence, and the people with policy influence rarely understand AI well. This award is for those who have worked to close that gap with verifiable effect.
AICI Algorithmic Justice Award 2024
Awarded to the individual or organisation that has done the most to identify, document, and contest AI systems causing harm to people — particularly people in communities with the least power to contest those harms. This award recognises that AI bias and discrimination are not hypothetical risks but documented harms, and that the researchers, advocates, and litigants who hold AI systems accountable are performing essential work that the broader AI industry systematically undervalues.
AICI Open AI Award 2024
Awarded to the model, tool, dataset, or research contribution released under an open licence that most expanded access to AI capability in 2024. AICI distinguishes open-weight from open-source: releasing model weights is valuable, but releasing training data, methodology, and evaluation frameworks is more so. This award recognises contributions that genuinely expand who can participate in AI development, not only who can download weights.
AICI AI in Science Award 2024
Awarded to the application of AI that produced the most significant scientific advance in 2024 — a result that accelerated, enabled, or fundamentally changed a domain of scientific inquiry. AlphaFold's impact on structural biology is the reference case: a problem that had challenged a field for 50 years, substantially resolved. AICI looks for AI contributions to science where the advance is measurable, the methodology is published, and the results are reproducible.