Paradigms of Intelligence Team
The Paradigms of Intelligence team (PAIR, later PI) is the Google research group Agüera y Arcas leads as Vice President and Fellow. The team's remit is unusually broad for industrial AI research: not to build better products but to investigate the fundamental nature of intelligence itself — as it arises in biological systems, in artificial architectures, and in the collective cognitive structures that emerge from their interaction. The team sits at the intersection of machine learning research, computational neuroscience, and the complexity science tradition associated with the Santa Fe Institute.
In the AI Story
The team's existence reflects Agüera y Arcas's conviction that the most important questions about AI are not engineering questions but scientific ones — what intelligence is, how it emerges, what architectures produce what capabilities, how biological and artificial forms of cognition relate. The framing is deliberate: paradigms (plural) of intelligence, not a single paradigm, and not artificial intelligence in isolation from natural kinds.
The research program has produced work on emergent capabilities in language models, on federated learning as a model of distributed cognition, on the relationship between biological and artificial neural architectures, and on the social implications of AI deployment. Team members publish in both technical venues and in broader intellectual publications like Noema, reflecting Agüera y Arcas's commitment to public intellectual engagement.
The team's position within Google is unusual. Most industrial AI research is product-adjacent, oriented toward capabilities that can be deployed. The Paradigms team operates at greater distance from product cycles, which affords the freedom to pursue longer-horizon questions but creates recurring pressure to justify the investment. Agüera y Arcas has defended the arrangement by arguing that foundational understanding of intelligence is the infrastructure on which every subsequent product depends.
The team's intellectual connections extend beyond Google through Agüera y Arcas's External Faculty appointment at the Santa Fe Institute and through formal collaborations with academic neuroscience and cognitive science groups. The hybrid structure — industrial resources, academic intellectual freedom — is itself a statement about what AI research ought to be.
Origin
The team evolved from earlier Google research groups Agüera y Arcas led after his 2014 arrival from Microsoft. The Paradigms of Intelligence framing was articulated progressively through the 2020s as the scope of the research broadened beyond specific product contributions to foundational questions.
Key Ideas
Paradigms, plural. The name insists that intelligence comes in multiple kinds, not a single paradigmatic form.
Foundational over applied. The team investigates the nature of intelligence rather than specific product capabilities.
Cross-disciplinary by design. Machine learning, neuroscience, complexity science, and philosophy of mind are all native disciplines to the team's work.
Public engagement as method. The team publishes in intellectual venues as well as technical ones, treating public understanding as part of the research mission.
Appears in the Orange Pill Cycle
Further reading
- Agüera y Arcas, Blaise. Google Research publications and the PI team's published work
- Mitchell, Melanie. Artificial Intelligence: A Guide for Thinking Humans (Farrar, Straus and Giroux, 2019)
- Santa Fe Institute, External Faculty and research program materials
- Agüera y Arcas, Blaise. Regular essays in Noema, 2020-present