When it comes to artificial intelligence, MIT and IBM were there at the beginning: laying the groundwork and creating some of the first programs – AI precursors – and theorizing how machines could become “intelligent.”
Today, collaborations like the MIT-IBM Watson AI Lab, which launched eight years ago, continue to provide expertise for the promise of tomorrow’s AI technology. This is important for the industries and labor force that stand to benefit, especially in the short term: from an estimated global economic benefit of $3-4 trillion and 80 percent productivity gains for knowledge workers and creative work, to significant incorporation of generic AI into business processes (80 percent) and software applications (70 percent) over the next three years.
While the industry has seen a surge in notable models, primarily in the last year, academia has continued to drive innovation, contributing most of the highly cited research. At the MIT-IBM Watson AI Lab, success takes the form of 54 patent disclosures, over 128,000 citations with an h-index of 162, and over 50 industry-driven use cases. The lab’s many accomplishments include improved stent placement with AI imaging techniques, reducing computational overhead, shrinking models while maintaining performance, and modeling interatomic potentials for silicate chemistry.
“The Lab is uniquely positioned to identify the ‘right’ problems to solve, which is what sets us apart from other institutions,” says Aud Oliva, director of the Lab at MIT and director of strategic industry engagement in the MIT Schwarzman College of Computing. “Furthermore, the experience our students gain from working on these challenges for enterprise AI translates to boosting their competitiveness in the job market and a competitive industry.”
“The MIT-IBM Watson AI Lab has made a tremendous impact by bringing together a rich set of collaborations between IBM and MIT researchers and students,” says Provost Ananth Chandrakasan, MIT co-chair of the lab and the Vannevar Bush Professor of Electrical Engineering and Computer Science. “By supporting cross-cutting research at the intersection of AI and many other disciplines, the Laboratory is advancing foundational work and accelerating the development of transformative solutions for our nation and the world.”
long-horizon work
As interest in AI continues to grow, many organizations are struggling to translate the technology into meaningful outcomes. A 2024 Gartner study found that, “At least 30% of generic AI projects will be abandoned after proof of concept by the end of 2025,” demonstrating ambition and widespread appetite for AI, but a lack of knowledge about how to develop and implement it to create immediate value.
Here, the laboratory shines, connecting research and deployment. The majority of the Lab’s current year research portfolio is designed to utilize and develop new features, capabilities, or products for IBM, the Lab’s corporate members, or real-world applications. The last of these include large language models, AI hardware, and foundation models, including multi-model, bio-medical, and geospatial. Inquiry-driven students and trainees are invaluable in this pursuit, providing enthusiasm and new perspectives while accumulating domain knowledge to help achieve and engineer advances in the field, as well as open new frontiers for exploration with AI as a tool.
The findings of the AAAI 2025 Presidential Panel on the Future of AI Research support the need for contributions from academia-industry collaborations such as laboratories in the AI field: “Academics have a role to play in providing independent advice and interpretation of these results. [from industry] And their results. The private sector focuses more on the short term, and universities and society focus more on the long-term perspective.
Bringing these strengths together, along with an emphasis on open sourcing and open science, can drive innovation that neither one can achieve alone. History shows that adopting these principles, and sharing code and making research accessible, has long-term benefits for both the field and society. Consistent with the missions of IBM and MIT, the Laboratory contributes technologies, findings, governance, and standards to the public domain through this collaboration, increasing transparency, accelerating reproducibility, and ensuring trustworthy progress.
The Lab was created to merge MIT’s deep research expertise with IBM’s industrial R&D capability, with the goal of driving breakthroughs in core AI methods and hardware, as well as new applications in areas such as health care, chemistry, finance, cybersecurity, and robust planning and decision making for business.
bigger is not always better
Today, larger foundation models are giving way to smaller, more task-specific models that offer better performance. Contributions from lab members like Song Han, associate professor in MIT’s Department of Electrical Engineering and Computer Science (EECS), and Chuang Gan of IBM Research, help make this possible through functions like Once-for-All and AWQ. Such innovations improve efficiency with better architectures, algorithm shrinkage, and activation-aware weight quantization, allowing models such as language processing to run on edge devices at faster speeds and lower latency.
As a result, foundation, vision, multimodal, and large language models have seen benefits, allowing the laboratory research groups of Oliva, MIT EECS Associate Professor Yoon Kim, and IBM Research members Rameshwar Panda, Yang Zhang, and Rogerio Ferris to move forward on the work. This includes techniques for filling models with external knowledge and the development of linear attention transformer methods for higher throughput than other state-of-the-art systems.
A gift in understanding and reasoning has also been observed in vision and multimodal systems. Works such as “Task2Sim” and “AdaFuse” demonstrate improved vision model performance if pre-trained on synthetic data, and how video action recognition can be enhanced by combining channels from past and present feature maps.
As part of a commitment to improving AI, MIT EECS Lab teams of engineering professor Gregory Wornell of Sumitomo Electric Industries, Chuang Gan of IBM Research, and David Cox, vice president of Foundational AI at IBM Research and IBM director of AI, have shown that model adaptability and data efficiency can go hand in hand. Two approaches, Evoscale and Chain-of-Action-Thought Reasoning (COAT), enable language models to make the most of limited data and computation by improving prior generation efforts through structured iteration, yielding better feedback. COAT uses meta-action frameworks and reinforcement learning to tackle logic-intensive tasks through self-improvement, while Evoscale brings a similar philosophy to code generation, developing high-quality candidate solutions. These technologies help enable resource-aware, targeted, real-world deployments.
“The impact of MIT-IBM research on our larger language model development efforts cannot be overstated,” says Cox. “We are seeing that smaller, more specialized models and tools are making a big impact, especially when they are combined. Innovations from the MIT-IBM Watson AI Lab help shape these technological directions and influence the strategies we take to market through platforms like Watsonx.”
For example, several lab projects have contributed features, capabilities, and uses to IBM’s Granite Vision, which, despite its compact size, provides impressive computer vision designed for document understanding. This comes at a time when there is a growing need for extraction, interpretation and reliable summarization of information and data contained in long formats for enterprise purposes.
The 2025 AAAI Panel concludes that AI and other achievements beyond direct research on a variety of topics are not only beneficial, but essential to advancing technology and uplifting society.
The lab’s Caroline Uhler and Devvrat Shah – both Andrew (1956) and Erna Viterbi professors at EECS and the Institute for Data, Systems, and Society (IDSS) – along with Kristjan Greenwald of IBM Research, go beyond the expertise. They are developing causal discovery methods to uncover how interventions affect outcomes, and identify which interventions achieve desired outcomes. Studies include developing a framework that can explain how “treatments” might work for different sub-populations, such as mobility restrictions on ecommerce platforms or on morbidity outcomes. The findings of this body of work could impact areas ranging from marketing and medicine to education and risk management.
“Advances in AI and other areas of computing are influencing how people create and tackle challenges in nearly every discipline. In the MIT-IBM Watson AI Lab, researchers recognize this cross-cutting nature of their work and its impact, interrogating problems from multiple perspectives and bringing real-world problems from industry to develop new solutions,” said MIT Lab co-chair, MIT Schwarzman College. says Dean of Computing and Henry Ellis Warren. (1894) Professor of Electrical Engineering and Computer Science.
A critical part of what enables this research ecosystem to flourish is the continued flow of student talent and their contributions through MIT’s Undergraduate Research Opportunities Program (UROP), the MIT EECS 6A Program, and the new MIT-IBM Watson AI Lab Internship Program. Overall, more than 70 young researchers have not only accelerated their technical skill development but, through the guidance and support of the lab’s mentors, have gained knowledge in the AI domain to become emerging practitioners. That’s why the Lab continuously strives to identify promising students at all stages of exploring the potential of AI.
“To unlock the full economic and social potential of AI, we need to foster ‘useful and efficient intelligence,’” says Sriram Raghavan, IBM Research VP for AI and president of IBM Labs. “To turn the promise of AI into progress, it is critical that we continue to focus on innovations to develop efficient, optimized, and fit-for-purpose models that can be easily adapted to specific domains and use cases. Academia-industry collaborations, such as the MIT-IBM Watson AI Lab, help advance the breakthroughs that make this possible.”
