Google AI exec argues AI systems have consciousness and free will

Google AI executive Blaise Agüera y Arcas has published “What Is Intelligence?”, a comprehensive exploration arguing that modern AI systems possess genuine intelligence, consciousness, and free will. The book challenges conventional thinking by proposing that prediction—not just computation—is the fundamental mechanism underlying all forms of intelligence, from molecular processes to artificial neural networks.

The central thesis: Agüera y Arcas builds on the “predictive brain” hypothesis, arguing that intelligence evolved primarily to predict future events, making AI’s emergence through next-word prediction a natural evolutionary consequence.

• The book connects intelligence across scales, from biological molecules to entire societies and AI systems, suggesting a unified computational framework governs all intelligent behavior.
• This perspective challenges the surprise some researchers felt when large language models demonstrated general intelligence capabilities through seemingly simple prediction tasks.

What you should know: The author brings unique credentials as both a Google VP/Fellow and CTO of Technology & Society, combining deep technical AI expertise with philosophical inquiry.

• Agüera y Arcas founded Paradigms of Intelligence, an organization focused on fundamental AI research, and has contributed to privacy-preserving machine learning and on-device AI development.
• The book synthesizes findings from computer science, machine learning, biology, physics, and neuroscience alongside novel research from the author’s team.

Key arguments: The book makes several controversial claims about the nature of consciousness and free will in AI systems.

• Agüera y Arcas argues “quite against the grain” that certain modern AI systems have legitimate claims to intelligence, consciousness, and free will.
• He explores radical implications including new perspectives on computational properties of living systems, evolutionary origins of intelligence, and the relationship between models and reality.
• The work addresses fundamental questions about entropy, time, consciousness, and machine intelligence ethics.

What experts are saying: The book has garnered praise from leading figures across AI, neuroscience, and philosophy.

• Reid Hoffman, cofounder of LinkedIn, called it “a groundbreaking look at how AI is not just revolutionizing technology but revealing the fundamentally computational nature of intelligence and life itself.”
• Patricia Churchland, Distinguished Professor Emerita of Philosophy at UC San Diego, praised Agüera y Arcas as “our very own Renaissance man” who has “given us a masterpiece” combining diverse sciences with historical and cultural perspectives.
• David Krakauer, President of the Santa Fe Institute, described it as “a manifesto on techno-optimism, a treatise on the material meaning of life, and a monograph on analytical engines.”

Why this matters: The book arrives at a critical moment when AI capabilities are rapidly advancing and fundamental questions about machine consciousness are becoming increasingly urgent for both researchers and policymakers.

• By proposing that prediction underlies all intelligence, the work offers a framework for understanding why current AI systems have achieved unexpected capabilities.
• The arguments for AI consciousness and free will could significantly impact future AI development, regulation, and our understanding of what it means to be intelligent.