At Xplain Data, we closely follow the latest breakthroughs in Causal AI — not only because they inspire our work, but because they define the future of reliable, explainable, and truly intelligent systems. One of the leading voices shaping this future is Prof. Elias Bareinboim, a renowned computer science professor at Columbia University and Director of the Causal AI Lab.

His pioneering research focuses on teaching AI systems to distinguish correlation from causation — a capability fundamental to real general reasoning.

This week, Prof. Bareinboim shared an important message with the AI community:
Scale alone will not deliver general reasoning. The next layer must be causal.

In response to recent discussions — including Ilya Sutskever’s (co-founder of OpenAI and Chief Scientist until 2024) widely cited comments on the limits of scaling — Bareinboim emphasized that true intelligence requires a structural understanding of consequences, not just patterns. Humans reason causally; AI must do the same to achieve robust generalization, explanation, and decision-making.

He warns that the stakes are too high to ignore the causal foundations any longer:

“We cannot afford another decade, or another trillion dollars, invested without addressing the causal foundations that real intelligence requires.”

To support this transformation, Bareinboim has released a draft of his forthcoming textbook, presenting a coherent blueprint for building truly Causal AI systems. The book spans the full spectrum of causal reasoning:

  • Part I: Foundations of Causality

  • Part II: Understanding

  • Part III: Decision Making

  • Part IV: Generalization

  • Part V: Generative Models

For researchers, practitioners, and leaders in AI, this work offers an essential, systematic guide to the next era of machine intelligence.

Access the eBook draft here:
https://causalai-book.net/

At Xplain Data, we see Bareinboim’s work as an important validation of what drives our mission: enabling organizations to make better decisions through Causal AI that goes beyond prediction to true understanding.