AAAI (the Association for the Advancement of Artificial Intelligence), founded in 1979 under the original name ‘American Association for Artificial Intelligence’ and renamed in 2007, is the principal non-profit scientific society devoted to advancing research in and responsible use of artificial intelligence. It is best known for its flagship annual peer-reviewed conference — one of the most selective and broadly scoped venues in AI — alongside the AI, Ethics and Society (AIES) conference, symposia, workshops, and the journal AI Magazine. AAAI also engages in education, public communication, and policy discussion concerning AI; administers the AAAI Fellows Program recognising sustained contributions to the discipline; and maintains historical continuity as the first professional society dedicated exclusively to AI, founded by leaders including Allen Newell, Edward Feigenbaum, Marvin Minsky, and John McCarthy. AAAI 2026 — the 40th annual conference — received 23,680 submissions and accepted 4,167 papers at a 17.6% acceptance rate, making it the most competitive edition in the conference’s history.
Semantic Classification
Content
Compositional Relationships (Components)
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:hasPart ai:AAAIConference))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:hasPart ai:PeerReview))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:hasPart ai:AIMagazine))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:hasPart ai:AAAIFellowsProgram))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:hasPart ai:Workshop))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:hasPart ai:Tutorial))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:hasPart ai:ProgrammeCommittee))Dependency Relationships
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:requires ai:PeerReview))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:requires ai:DoubleBlindReview))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:dependsOn ai:AIResearch))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:dependsOn ai:AcademicCommunity))Capability Relationships
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:enables ai:ResearchDissemination))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:enables ai:BenchmarkStandard))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:enables ai:AIForSocialGood))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:enables ai:AIGovernance))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:enables ai:AIAlignment))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:enables ai:AIPolicy))Implementation Relationships
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:implements ai:PeerReviewProcess))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:implements ai:FellowsRecognitionProcess))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:supports ai:ArtificialIntelligence))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:supports ai:MachineLearning))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:supports ai:ResponsibleAI))Reduction Relationships
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:reducesTo ai:AcademicConference))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:reducesTo ai:ScientificSociety))
SubClassOf(ai:AAAI
ObjectSomeValuesFrom(ai:reducesTo ai:ResearchCommunity))About
AAAI occupies a distinctive institutional position in the global AI ecosystem as both a scientific society with membership-based governance and a premier peer-reviewed conference series with a breadth that distinguishes it from all ML-focused alternatives. Where NeurIPS is anchored in neural computation and statistical learning and ICML maintains a strong grounding in learning theory, AAAI’s historical roots in Symbolic AI and Knowledge Representation give it a wider disciplinary scope that has proven resilient across multiple paradigm shifts in AI: from the expert-systems boom and bust of the 1980s, through the probabilistic and statistical learning revolutions of the 1990s and 2000s, to the deep learning era of the 2010s, and now into the era of Large Language Model, Agentic AI, and Multimodal AI. This breadth makes AAAI papers a particularly useful signal of AI progress across the full technical stack — from foundational theory and Reasoning systems through applied Machine Learning and sociotechnical concerns including AI Ethics, Fairness in Machine Learning, and Explainability.
The organisation functions as a true scientific society rather than merely a conference management entity. It publishes AI Magazine, a quarterly peer-reviewed popular-academic hybrid publication that has served as the field’s principal practitioner-accessible journal since 1980 and carries authoritative survey and perspective articles. The AAAI Fellows Program, established in 1990, recognises individuals who have made “significant, sustained contributions to the field of artificial intelligence” — a historically small and selective class that includes many of AI’s most consequential figures. Notable AAAI Fellows include Andrew Ng, Yann LeCun, Yoshua Bengio, Geoffrey Hinton, Judea Pearl, Barbara Grosz, and Peter Norvig. The 2024 and 2025 cohorts included Cynthia Rudin (Duke, for interpretable ML and trustworthy AI), Pascale Fung (HKUST, for conversational AI and ethical AI principles), Lynne E. Parker (University of Tennessee, for distributed robotics and AI policy leadership), and David Silver (Google DeepMind, London, for deep reinforcement learning and game-playing AI). AAAI also administers the Classic Paper Award (for papers ten or more years old whose influence has proven fundamental) and the Outstanding Paper Award (for each annual conference’s most significant contributions).
The AAAI/ACM Conference on AI, Ethics, and Society (AIES), co-organised with ACM since 2018, is the field’s leading dedicated venue for interdisciplinary research on AI’s societal implications, bringing together computer scientists, ethicists, legal scholars, social scientists, and policy researchers. AIES 2024 was held in San Jose, California; AIES 2025 continued the series with an expanded track on AI Governance and policy.
Components / Architecture
The AAAI annual conference comprises several formally distinct programme tracks:
-
Main Technical Track — The primary peer-reviewed programme, accepting original research across the full spectrum of AI including Knowledge Representation, Planning and Scheduling, Reasoning, Machine Learning, Deep Learning, Natural Language Processing, Computer Vision, Robotics, Multi-Agent System research, Constraint Satisfaction, search, and Causal Inference. Submissions undergo Double-Blind Review by a Programme Committee of area chairs and reviewers; accepted papers are presented as oral talks or poster presentations. In AAAI 2026 (the 40th conference), 23,680 submissions were reviewed and 4,167 accepted at a 17.6% acceptance rate — approximately twice the submission volume of AAAI 2025 (12,957 valid submissions, 3,032 accepted at 23.4%).
-
Special Track on AI for Social Impact — A dedicated track for research demonstrating measurable positive societal applications of AI. AAAI 2026 awarded 2 Outstanding Papers in this track, covering slum detection for urban planning and plant trait mapping for agricultural sustainability.
-
Special Track on AI Alignment — Introduced at AAAI 2026 to reflect the community’s growing focus on making AI systems controllable, interpretable, and safe. This track spans AI Safety, value alignment, reward modelling, and human-AI collaboration research.
-
AAAI/ACM AIES — The joint ethics conference, run annually alongside or adjacent to the main conference, providing a dedicated venue for interdisciplinary AI ethics and governance research at the boundary of Responsible AI and AI Policy.
-
Workshop Programme — Dozens of focused Workshop events organised by community members, covering emerging sub-fields and specialised applications. Workshops serve as incubators for research that appears in later main-track submissions.
-
Tutorial Programme — Half-day and full-day Tutorial sessions by domain experts, covering both foundational methods and state-of-the-art developments in AI.
-
Doctoral Consortium — A structured mentorship programme for PhD candidates in AI to present and receive feedback on dissertation work from senior researchers, supporting the pipeline of future AI researchers.
-
AI Magazine — AAAI’s peer-reviewed practitioner journal, publishing survey articles, perspective pieces, and accessible technical articles since 1980. Functions as the de facto house journal of the professional AI community.
-
AAAI Fellows Program — Annual recognition of up to 5–8 individuals per year who have made sustained, significant contributions to AI. Fellows are nominated by members and elected by the existing Fellows body.
Use Cases / Major Families
AAAI serves distinct community functions:
Breadth-First Research Publication — AAAI accepts research across the full AI technical stack in a single venue, making it the conference of choice for researchers working at the intersection of sub-fields (e.g., combining Knowledge Representation with Machine Learning, or Planning and Scheduling with Reinforcement Learning). NeurIPS and ICML tend to favour more narrowly focused ML contributions; IJCAI is a broadly comparable alternative but held biennially.
AI Ethics and Policy Research Dissemination — Via the AIES track and the main conference’s ethics-oriented area, AAAI is the primary peer-reviewed venue for AI Ethics, Fairness in Machine Learning, Explainability, Responsible AI, and AI Governance research, complementing the more technically focused ethics work appearing at NeurIPS and FAccT.
Professional Society Functions — AAAI’s membership structure, Fellows recognition, awards programme, and AI Magazine publication provide the professional-society functions that NeurIPS (a foundation) and ICML (organised by IMLS) also provide but in more limited forms.
Symbolic and Classical AI Preservation — AAAI remains the principal venue where Symbolic AI, Knowledge Representation and Reasoning, constraint programming, and Planning and Scheduling communities publish their premier work, preventing these disciplines from being marginalised by the dominance of Deep Learning in other venues.
Early-Stage AI Safety Research — The Special Track on AI Alignment introduced in 2026 positions AAAI as an important outlet for AI Safety and alignment research, complementing the dedicated workshops at NeurIPS and ICML and the AI Safety Research conference (SaTML).
Policy and Standards Input — AAAI’s presidential panels, symposia, and ethics conference inform public discourse and regulatory deliberations. AAAI representatives have testified to the US Congress, contributed to NIST AI Risk Management Framework development, and engaged with EU AI Act consultations.
Academic Context
AAAI traces its intellectual lineage to the foundational era of AI. Allen Newell — the first AAAI president — and Herbert Simon had demonstrated General Problem Solver (GPS) as early as 1957 as a prototype of human-like problem solving through symbolic manipulation. The first AAAI conference (AAAI-80) took place in 1980 and brought together the growing US AI research community that had been working on Symbolic AI systems including expert systems, Knowledge Representation languages (KL-ONE, frame systems), and natural language understanding.
Key academic figures associated with AAAI leadership and publication include: Marvin Minsky (MIT, frames and the society of mind); Edward Feigenbaum (Stanford, expert systems); John McCarthy (Stanford, LISP, situation calculus); Judea Pearl (UCLA, Bayesian networks, causality — Turing Award 2011); Barbara Grosz (Harvard, discourse structure); Stuart Russell and Peter Norvig (authors of the field’s canonical textbook “Artificial Intelligence: A Modern Approach”); and Yann LeCun, Geoffrey Hinton, and Yoshua Bengio (the trio awarded the 2018 Turing Award for deep learning).
The AAAI proceedings archive constitutes a comprehensive longitudinal record of AI research progress from 1980 to the present, making it a unique resource for meta-scientific analysis of the field’s evolution. The shift in AAAI’s paper distribution from predominantly Symbolic AI and search methods (pre-2010) through probabilistic and graphical model approaches (2000s) to Deep Learning (2015+) and Large Language Model and Agentic AI topics (2023+) mirrors and documents the paradigm transitions of the discipline.
The journal AI Magazine has published landmark survey articles including Nilsson’s “Eye on the Prize” (2003), Pearl’s retrospectives on causality, and survey articles on deep learning, Reinforcement Learning, and AI safety, forming an accessible intellectual history of the field.
Current Landscape (2026)
AAAI 2026 — the 40th edition, held in Singapore January 20–27, 2026 — established several records. The 23,680 paper submissions represent approximately an 83% increase over AAAI 2025’s 12,957 valid submissions, driven primarily by the explosion of interest in Large Language Model, Retrieval-Augmented Generation, Agentic AI, and Multimodal AI research since the 2023 GPT-4 release. The accepted paper count of 4,167 (17.6% acceptance rate) was deliberately held lower in percentage terms than prior years to maintain quality standards despite the submission surge.
The conference theme “Creating Collaborative Bridges Within and Beyond AI” operationalised a community consensus that the research agenda must move beyond raw capability scaling toward AI Safety, interpretability, controllability, and alignment. Five Outstanding Papers were awarded in the Main Technical Track spanning causal discovery, vision-language models, robotic manipulation, and other areas. The Special Track on AI for Social Impact awarded 2 Outstanding Papers for research on slum detection and plant trait mapping. The new Special Track on AI Alignment formalised the safety and alignment research strand that had previously been represented across several workshops and symposia.
AAAI’s AAAI Fellows Program welcomed new members at a dedicated dinner at AAAI-26, with the 2025 cohort (to be announced December 2025) representing the first class whose work substantially includes Large Language Model and Agentic AI contributions alongside traditional AI research areas.
The AAAI/ACM AIES conference continues to grow as regulatory frameworks (EU AI Act, US Executive Orders on AI) increase demand for academic AI ethics and governance research that can inform policy.
UK Context
The UK AI research community has maintained a significant presence at AAAI throughout its history and increasingly so in the 2020s as the UK’s national AI investment has grown. Several observations characterise the UK relationship with AAAI:
University of Edinburgh — The School of Informatics is the UK’s leading AI research centre and maintains a strong AAAI publication record across Natural Language Processing, Multi-Agent System research, Planning and Scheduling, and increasingly Large Language Model and Responsible AI research. Edinburgh researchers have served on AAAI Programme Committee and area chair roles. The Edinburgh-based Centre for Reasoning, Language and Agents (CRLA) produces work directly relevant to AAAI’s core programme areas.
Imperial College London — Imperial’s AI research group publishes across Machine Learning, Computer Vision, and Robotics at AAAI. The Dyson School of Design Engineering and the Department of Computing contribute work on human-AI interaction and Explainability that aligns with AAAI’s breadth mandate. Imperial, Cambridge, and Oxford have jointly launched QRT Labs (2026) for multidisciplinary AI and mathematics research with direct relevance to AAAI topics.
University of Oxford — Oxford’s Department of Computer Science, Future of Humanity Institute (renamed Existential Risk Observatory), and Oxford Internet Institute contribute AI Safety, AI Governance, and Reasoning research to AAAI. The Leverhulme Centre for the Future of Intelligence (Cambridge/Oxford partnership) produces AI policy and ethics research published at AAAI’s AIES track.
University of Cambridge — The Cambridge Computer Lab and the Leverhulme CFI contribute work spanning Causal Inference, AI Ethics, and formal verification of AI systems. Cambridge hosts the AI Safety Research group whose work appears at both NeurIPS safety workshops and AAAI.
UCL — University College London, which leads the UKRI national generative AI hub, contributes Deep Learning, Reinforcement Learning, and Generative AI research to AAAI alongside its dominant NeurIPS publication presence. UCL’s involvement in the national generative AI hub (with Imperial, Cardiff, Cambridge, Oxford, Manchester, Edinburgh) will generate work spanning multiple AAAI programme areas.
DeepMind (London) — Google DeepMind’s UK research labs, while primarily publishing at NeurIPS and ICML, submit foundational Reinforcement Learning, Multi-Agent System, and safety research to AAAI. David Silver (DeepMind) became an AAAI Fellow in recognition of contributions to deep reinforcement learning.
Northern England — The universities of Manchester, Leeds, Sheffield, and Newcastle collectively contribute AI research to AAAI across applied Machine Learning, Natural Language Processing, and intelligent systems. The UKRI AI hub at Manchester focuses on AI for healthcare and has submitted work to AAAI’s AI for Social Impact track.
Future Directions (2026-2030)
AAAI’s trajectory over the next four years is shaped by several convergent pressures:
Scale Management — The jump from ~13,000 to ~24,000 submissions in a single year poses severe Programme Committee scaling challenges. AAAI is exploring: AI-assisted review assignment and reviewer matching; tiered reviewing pipelines with desk rejection; and expanded area chair capacity. The Open Review transparency question — whether to publish reviews publicly as ICLR does — is under active debate within the community.
AI Safety and Alignment Integration — The new AI Alignment special track at AAAI 2026 is expected to grow into a standalone conference track with dedicated area chairs. AAAI is positioned to become the primary annual peer-reviewed venue for AI Safety research, complementing the safety-workshop ecosystem at NeurIPS and ICML with a more formal acceptance process.
Interdisciplinary Expansion — The AIES conference is expanding its scope to cover AI Governance frameworks in response to the EU AI Act and equivalent national regulations. AAAI is increasing co-organisation with law, social science, and ethics societies to maintain its mandate as the premier interdisciplinary AI venue.
Large Language Model and Agentic AI Absorption — As Large Language Model and Agentic AI research continues to dominate AI publication, AAAI must create dedicated programme tracks for these areas while preventing them from crowding out foundational AI subfields (Knowledge Representation, Planning and Scheduling, Reasoning) that remain core to AAAI’s mission.
Global Rotation — Following AAAI 2026 in Singapore, the conference is expected to continue its periodic rotation through North America, Europe, and Asia, reflecting the global distribution of the AI research community and the declining dominance of US institutions in AI publication counts.
AI Policy Influence — As governments worldwide implement AI regulations informed by scientific evidence, AAAI’s role as a convener of AI policy-relevant research — through AIES, presidential panels, and fellowship recognition for policy leadership — is expected to grow. AAAI is positioning itself as a primary academic partner for regulatory bodies including NIST, the UK AI Safety Institute, and the EU AI Office.
Research & Literature
- AAAI. (1979–2026). Proceedings of the AAAI Conference on Artificial Intelligence (Vols. 1–40). AAAI Press. https://aaai.org/aaai-publications/aaai-conference-proceedings/
- AAAI. (2026). The 40th Annual AAAI Conference on Artificial Intelligence: AAAI-26. https://aaai.org/conference/aaai/aaai-26/
- AAAI. (2026). AAAI 2026 Accepted Papers: Results and Statistics. https://aaai.org/conference/aaai/aaai-26/main-technical-track/
- Bohrium Research. (2026). “AAAI 2026 Accepted Papers: Outstanding Papers, Research Trends & Full Highlights.” https://www.bohrium.com/en/blog/research-notes/aaai-2026-accepted-papers-highlights/
- AAAI. (2026). AAAI-26 Research Spotlights from Singapore. https://aaai.org/conference/aaai/aaai-26/research-spotlights/
- AAAI. (2026). AAAI Conference Paper Awards and Recognition. https://aaai.org/about-aaai/aaai-awards/aaai-conference-paper-awards-and-recognition/
- AAAI. (2025). About AAAI. https://aaai.org/about-aaai/
- Wikipedia. (2026). “Association for the Advancement of Artificial Intelligence.” https://en.wikipedia.org/wiki/Association_for_the_Advancement_of_Artificial_Intelligence
- Hayes-Roth, F., Waterman, D. A., & Lenat, D. B. (Eds.). (1983). Building Expert Systems. Addison-Wesley. [Foundational AAAI-era text]
- Russell, S., & Norvig, P. (2022). Artificial Intelligence: A Modern Approach (4th ed.). Pearson. [The canonical AI textbook, reflecting AAAI’s intellectual scope]
- Nilsson, N. J. (1998). Artificial Intelligence: A New Synthesis. Morgan Kaufmann. [Classic AAAI-community reference]
- Pearl, J. (2000). Causality: Models, Reasoning, and Inference. Cambridge University Press. [Turing Award 2011; key AAAI research line]
- McCarthy, J. (2007). “A History of AAAI.” AI Magazine, 28(4). [The founding president’s retrospective]
- CRA. (2025). “CRA Congratulates New AAAI, ACM, and IEEE Fellows.” https://cra.org/crn/2025/03/cra-congratulates-new-aaai-acm-and-ieee-fellows-and-acm-distinguished-members/
- AAAI. (2025). The AAAI Fellows Program. https://aaai.org/about-aaai/aaai-awards/the-aaai-fellows-program/
- AAAI. (2025). Artificial Intelligence, Ethics, and Society (AIES). https://aaai.org/conference/aies/
- ACM Conferences. (2024). Proceedings of the 2024 AAAI/ACM Conference on AI, Ethics, and Society. https://dl.acm.org/doi/proceedings/10.5555/3716662
- Newell, A., & Simon, H. A. (1976). “Computer Science as Empirical Inquiry: Symbols and Search.” Communications of the ACM, 19(3), 113–126. [Turing Award lecture; seminal for AAAI’s intellectual foundation]
- Minsky, M. (1975). “A Framework for Representing Knowledge.” In P. H. Winston (Ed.), The Psychology of Computer Vision. McGraw-Hill. [Frame systems; Knowledge Representation foundation]
- Brachman, R. J., & Schmolze, J. G. (1985). “An Overview of the KL-ONE Knowledge Representation System.” Cognitive Science, 9(2), 171–216. [Canonical Knowledge Representation paper from early AAAI era]
- Heckerman, D. (1997). “Bayesian Networks for Data Mining.” Data Mining and Knowledge Discovery, 1(1), 79–119. [AAAI-community probabilistic Reasoning contribution]
- Shoham, Y., & Leyton-Brown, K. (2009). Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations. Cambridge University Press. [Multi-Agent System research; prominent AAAI sub-field]
- Lifschitz, V., Brewka, G., & Hertzberg, J. (Eds.). (2008). Logic Programming, Knowledge Representation, and Nonmonotonic Reasoning. Springer. [Knowledge Representation and Reasoning domain]
- Ghallab, M., Nau, D., & Traverso, P. (2004). Automated Planning: Theory and Practice. Morgan Kaufmann. [Planning and Scheduling canonical text; core AAAI sub-field]
- AIES Conference. (2025). Call for Papers — AIES 2025. https://www.aies-conference.com/2025/call-for-papers/
- Oxford University. (2025). “Oxford Institute for Ethics in AI Launches Accelerator Fellowship Programme.” https://www.ox.ac.uk/news/2025-03-03-oxford-institute-ethics-ai-launches-accelerator-fellowship-programme
- RIKEN AIP. (2026). “Nine Papers Accepted to AAAI-26.” https://aip.riken.jp/news/aaai26/
- 36KR. (2026). “AAAI 2026 Results Announced: 17.6% Acceptance Rate for 23,000 Submissions.” https://eu.36kr.com/en/p/3546803868545153