Modern artificial intelligence traces its roots back to the summer of 1956, when a small group of researchers gathered at Dartmouth College to explore whether machines could be made to think. The monthlong meeting, later known as the Dartmouth Summer Research Project on Artificial Intelligence, brought together prominent engineers and mathematicians to debate ideas such as machine language, abstraction, and learning. Buoyed by early successes in digital computing, optimism ran high. Within a decade, some pioneers confidently predicted that human-level artificial intelligence was just a generation away.
Reality proved far more complex.
While computers grew faster and more powerful, the dream of building machines with human-like intelligence remained elusive. In recent years, however, machine learning has fueled major breakthroughs. AI systems can now recognize speech, translate languages, diagnose medical images, and outperform human champions in games such as Jeopardy. Yet these advances, according to MIT professor Tomaso Poggio, highlight how far the field still has to go.
“Despite impressive progress, we still don’t understand how intelligence actually emerges in the brain,” Poggio says. “And without that understanding, we can’t build machines that truly match human intelligence across a wide range of tasks.”
A Return to Big Questions
Poggio believes artificial intelligence research needs to reclaim its original, ambitious vision. Rather than focusing solely on narrowly specialized systems, he argues for a deeper integration of neuroscience, cognitive science, and computer science.
Also Read – Artificial Intelligence in Daily Life: What U.S. Data Reveals About Public Views
“We now know vastly more about how biological brains work than we did decades ago,” Poggio says. “That knowledge gives us an opportunity to rethink how we design intelligent machines.”
That philosophy is now being backed by significant federal support. The National Science Foundation has announced funding for the Center for Brains, Minds and Machines (CBMM), a new interdisciplinary research center based at MIT and led by Poggio. The center will receive $25 million over five years as part of the NSF’s Science and Technology Centers Integrative Partnerships program.
From Campus Vision to Global Collaboration
CBMM grew out of the MIT Intelligence Initiative, an effort launched in 2011 to bring together researchers studying intelligence from different perspectives. The idea gained momentum after MIT leadership encouraged faculty to pursue bold, cross-disciplinary projects during the university’s 150th anniversary.
Working with Joshua Tenenbaum, a cognitive scientist and computer scientist at MIT, Poggio helped shape a program that bridged the Department of Brain and Cognitive Sciences and the Department of Electrical Engineering and Computer Science. The goal was simple but ambitious: understand intelligence as a unified phenomenon rather than a collection of isolated skills.
Today, CBMM reflects that same collaborative spirit. While headquartered at MIT, the center includes faculty from institutions such as Harvard, Stanford, Cornell, UCLA, Rockefeller University, and the Allen Institute for Brain Science. International partners span Europe, Asia, and the Middle East, while industry collaborators include leading technology and robotics companies such as Google, Microsoft, IBM, DeepMind, and Boston Dynamics.
Students and postdoctoral researchers supported by the center will be encouraged to work across disciplines, often with multiple advisors from different fields.
Four Pillars of Intelligence Research
CBMM’s research agenda is organized around four interconnected themes.
The first focuses on integrating core components of intelligence, including vision, language, and motor control. Unlike many AI systems that treat these abilities separately, the center aims to understand how they work together in humans.
The second theme examines the neural circuits underlying intelligence, combining insights from neurobiology with engineering and computational modeling.
Also Read – New Artificial Intelligence Institute Boosts Clemson University’s Edge in a Competitive Tech Field
A third area explores how intelligence develops over time, particularly in children. Researchers are interested in how interaction with the environment shapes learning and cognition, and how early experiences influence lifelong capabilities.
The final theme centers on social intelligence — how humans understand, communicate with, and learn from one another.
Supporting all four areas is a theoretical framework designed to connect experimental findings with computational models.
Intelligence as a Story, Not a Snapshot
Patrick Winston, MIT professor and research coordinator for CBMM, says one of the center’s goals is to move beyond surface-level pattern recognition. Humans, he explains, don’t just see objects — they interpret situations.
For example, a person can easily tell the difference between someone drinking from a glass and someone raising it in a toast, even though the images may look similar. At the same time, humans recognize that a person sipping water and a cat lapping from a faucet are both performing the same action: drinking.
“What matters is the underlying story,” Winston says. “Humans understand actions and intentions, not just visual features.”
Development plays a crucial role as well. Research shows that without proper sensory stimulation early in life, key cognitive abilities may never fully develop. According to Winston, intelligence emerges through continuous interaction between the brain and the environment, with different systems — vision, language, and motor skills — shaping one another from the very beginning.
Building Smarter Machines by Understanding Ourselves
By uniting insights from neuroscience, psychology, and computer science, CBMM hopes to close the gap between artificial and natural intelligence. Rather than chasing quick technological wins, the center aims to answer fundamental questions about how intelligence works — and how it might one day be recreated in machines.
For Poggio and his collaborators, the message is clear: the future of artificial intelligence depends not just on better algorithms or faster computers, but on a deeper understanding of the human mind itself.
