AI@50, formally known as the Dartmouth Artificial Intelligence Conference: The Next Fifty Years, was a conference held from July 13 to July 15, 2006, at Dartmouth College in Hanover, New Hampshire. Organized by philosopher James H. Moor, the event marked the 50th anniversary of the 1956 Dartmouth workshop, which is widely regarded as the founding moment of artificial intelligence as a distinct research discipline. The conference brought together leading researchers to assess the progress of AI over its first half-century and to chart future challenges and breakthroughs.
Five of the ten original attendees of the 1956 Dartmouth workshop were present at AI@50: Marvin Minsky, Ray Solomonoff, Oliver Selfridge, Trenchard More, and John McCarthy. Their participation provided a direct link between the field's origins and its contemporary state. The conference was sponsored by Dartmouth College, General Electric, and the Frederick Whittemore Foundation. A $200,000 grant from the Defense Advanced Research Projects Agency (DARPA), awarded to George Cybenko, required a report of the proceedings that would analyze progress on AI's original challenges, document participants' views on future research and development challenges, and relate those challenges to trends in related fields such as control theory, signal processing, information theory, statistics, and optimization theory. Moor, serving as conference director, published a summary report in AI Magazine.
Historical Context
The 1956 Dartmouth workshop, organized by John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon, is commonly cited as the birthplace of AI. Over a six-week period, participants explored topics such as neural networks, problem-solving, and creativity, laying the groundwork for decades of research. By 2006, AI had evolved through periods of optimism, funding cuts (the so-called "AI winters"), and resurgences driven by advances in machine learning, neural networks, and increased computational power. AI@50 sought to reflect on this trajectory and to look forward to the next fifty years.
Conference Program and Themes
The conference was structured into themed sessions, each featuring talks by prominent researchers. The program covered a wide range of topics, from the philosophical foundations of thought to practical applications in vision, language, and games.
AI: Past, Present, Future
John McCarthy, a pioneer of the field, delivered a talk titled "What Was Expected, What We Did, and AI Today," reflecting on the gap between early expectations and actual achievements. Marvin Minsky presented "The Emotion Machine," discussing his then-recent book on how emotions and common sense might be integrated into AI systems.
The Future Model of Thinking
Ron Brachman and Hector Levesque addressed "A Large Part of Human Thought," focusing on knowledge representation and reasoning. David Mumford, a mathematician, posed "What is the Right Model for 'Thought'?" exploring connections between statistics and cognition. Stuart Russell discussed "The Approach of Modern AI," emphasizing probabilistic methods and rational decision-making.
The Future of Network Models
Geoffrey Hinton and Simon Osindero presented "From Pandemonium to Graphical Models and Back Again," tracing the evolution of connectionist ideas. Rick Granger spoke on "From Brain Circuits to Mind Manufacture," linking neuroscience to AI design.
The Future of Learning & Search
Oliver Selfridge, one of the original Dartmouth attendees, discussed "Learning and Education for Software: New Approaches in Machine Learning." Ray Solomonoff, another original participant, presented "Machine Learning - Past and Future," reflecting on his foundational work in algorithmic probability. Leslie Pack Kaelbling spoke on "Learning to be Intelligent," and Peter Norvig, then director of research at Google, gave a talk titled "Web Search as a Product of and Catalyst for AI," highlighting the symbiotic relationship between search engines and AI research.
The Future of AI
Rod Brooks, known for his work in robotics, delivered "Intelligence and Bodies," arguing for the importance of embodiment. Nils Nilsson presented "Routes to the Summit," outlining pathways to human-level AI. Eric Horvitz discussed "In Pursuit of Artificial Intelligence: Reflections on Challenges and Trajectories," drawing on his experience at Microsoft Research.
The Future of Vision
Eric Grimson spoke on "Intelligent Medical Image Analysis: Computer Assisted Surgery and Disease Monitoring," showcasing applications in healthcare. Takeo Kanade presented "Artificial Intelligence Vision: Progress and Non-Progress," assessing both successes and persistent difficulties. Terry Sejnowski gave "A Critique of Pure Vision," questioning the sufficiency of vision-only approaches.
The Future of Reasoning
Alan Bundy addressed "Constructing, Selecting and Repairing Representations of Knowledge," focusing on automated reasoning. Edwina Rissland discussed "The Exquisite Centrality of Examples," emphasizing case-based reasoning. Bart Selman spoke on "The Challenge and Promise of Automated Reasoning," highlighting advances in SAT solvers and their impact.
The Future of Language and Cognition
Trenchard More, another original Dartmouth participant, presented "The Birth of Array Theory and Nial," discussing his work on array-based programming. Eugene Charniak gave a talk titled "Why Natural Language Processing is Now Statistical Natural Language Processing," reflecting the field's shift toward probabilistic methods. Pat Langley discussed "Intelligent Behavior in Humans and Machines," advocating for cognitive systems research.
The Future of the Future
Ray Kurzweil delivered "Why We Can Be Confident of Turing Test Capability Within a Quarter Century," predicting rapid progress toward human-level AI. George Cybenko presented "The Future Trajectory of AI," and Charles J. Holland offered "DARPA's Perspective," outlining the agency's research priorities.
AI and Games
Jonathan Schaeffer, known for his work on the checkers program Chinook, spoke on "Games as a Test-bed for Artificial Intelligence Research." Danny Kopec discussed "Chess and AI," and Shay Bushinsky presented "Principle Positions in Deep Junior's Development," detailing the chess program that drew a match against world champion Garry Kasparov in 2003.
Future Interactions with Intelligent Machines
Daniela Rus, now a leader in robotics, presented "Making Bodies Smart," discussing modular robots. Sherry Turkle gave "From Building Intelligences to Nurturing Sensibilities," exploring the psychological and social dimensions of human-machine interaction.
Selected Submitted Papers
The conference also included sessions for submitted papers. In "Future Strategies for AI," J. Storrs Hall presented "Self-improving AI: An Analysis," Selmer Bringsjord offered "The Logicist Manifesto," Vincent C. Müller asked "Is There a Future for AI Without Representation?," and Kristinn R. Thórisson discussed "Integrated A.I. Systems." In "Future Possibilities for AI," Eric Steinhart presented "Survival as a Digital Ghost," Colin T. A. Schmidt asked "Did You Leave That 'Contraption' Alone With Your Little Sister?," Michael Anderson and Susan Leigh Anderson discussed "The Status of Machine Ethics," and Marcello Guarini addressed "Computation, Coherence, and Ethical Reasoning."
Significance and Legacy
AI@50 served as a milestone reflection on the state of AI. The presence of five original Dartmouth participants underscored the field's relatively short but eventful history. The conference's emphasis on future challenges, including the need for breakthroughs in reasoning, learning, and interaction, anticipated many developments that followed, such as the rise of deep learning and large language models. The DARPA-funded report contributed to ongoing discussions about AI's trajectory and its relationship to other scientific disciplines.
Videos of the talks were made available online, and the official conference website preserved the program and abstracts. Moor's summary report in AI Magazine remains a key reference for historians of AI. The conference is often cited in retrospectives as a moment when the community took stock of its achievements and charted a course for the future.
References
- Moor, James. "The Dartmouth College Artificial Intelligence Conference: The Next Fifty Years." AI Magazine 27:4 (2006): 87-91.
- Official conference website: Dartmouth Artificial Intelligence Conference: The Next Fifty Years.
- Norvig, Peter. "Pictures from AI@50." Photographs of conference presenters.