The future of intelligence | Demis Hassabis (Co-founder and CEO of DeepMind)

The future of intelligence | Demis Hassabis (Co-founder and CEO of DeepMind)

by Demis Hassabis

Formal & Physical Sciences AI — Models & Research UYQArtificial intelligence (AI)UYQDArtificial general intelligence (AGI)
🎙 Demis Hassabis 👥 911K 📅 December 16, 2025 ⏱ 56 min 👁 10.6M 📄 expert opinion 🧭 2026-08-02
Available in: English (current) Français

Keywords

AGIworld modelsscalingAI safetyAlphaFold

Summary

In this interview, Demis Hassabis discusses the rapid progress in AI over the past year, highlighting advances in multimodal models like Gemini 3 and the emergence of world models. He emphasizes the importance of solving ‘root node’ problems such as fusion energy and material science, which could unlock transformative benefits. Hassabis addresses the ‘jagged intelligence’ of current AI systems, noting their inconsistency across tasks, and outlines the need for better reasoning and continual learning. He reflects on the balance between scientific research and commercialization, acknowledging the pressures of the AI race. He shares his vision for AGI, including the role of simulations and world models, and discusses the importance of building ethical AI. The conversation also touches on the potential of AI to accelerate scientific discovery, the concept of Turing machines, and the challenges of leading a major AI lab. Throughout, Hassabis maintains a thoughtful and measured tone, emphasizing the need for both scaling and innovation to achieve AGI.

159 words

Critical Evaluation

This interview provides a valuable and insightful perspective on the future of AI from one of the field’s leading figures. Demis Hassabis demonstrates a deep understanding of both the technical and societal aspects of AI development. His discussion of ‘jagged intelligence’ is particularly illuminating, as it highlights the uneven capabilities of current systems and the challenges that remain. The emphasis on ‘root node’ problems, such as fusion and material science, shows a long-term vision that extends beyond commercial applications. Hassabis’s candid reflection on the trade-offs between scientific rigor and commercial pressures is refreshing and adds credibility to his statements. The conversation is well-structured, with Hannah Fry asking probing questions that allow Hassabis to elaborate on complex topics. The sources cited are primarily internal projects and partnerships, which are credible but not independently verified. The title accurately reflects the content, and the interview avoids hyperbole, offering a balanced and realistic assessment of AI’s trajectory. Overall, this is a high-quality discussion that provides valuable insights for anyone interested in the future of AI.

171 words

Title / Content Match

The title accurately reflects the content, which focuses on the future of AI and intelligence as discussed by Demis Hassabis.

Quality & Reliability

9/10

High credibility due to Demis Hassabis's leadership at DeepMind, balanced views, and references to concrete projects. However, it's an interview with opinions and forward-looking statements, not peer-reviewed research.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

Concurring Sources

Contribution & Novelties

This interview provides a unique insider perspective on the strategic direction of Google DeepMind, including the emphasis on world models and the concept of ‘jagged intelligence.’ It offers a balanced view on the scaling debate and the importance of scientific rigor. The discussion on AGI and the role of simulations is particularly insightful.

Pour aller plus loin :

  • World Models — Background on the concept of world models in AI.
  • AlphaFold — Overview of the protein structure prediction breakthrough.
  • Turing Machine — Foundational concept in computability theory.

87 words

Radar Profile

The radar profile shows high scores across all dimensions, indicating a well-rounded and reliable content. The lowest score is in technical depth, but it remains strong, reflecting the accessible yet informative nature of the interview.

Reliability 9/10

💬 Très positif. Sur les 30 commentaires analysés, le public exprime une admiration et une confiance marquées envers Demis Hassabis, le décrivant comme honnête, passionné et crédible, et saluant son approche scientifique et éthique.