Faut-il laisser l'IA décider à notre place ? Un ex-McKinsey répond

Faut-il laisser l'IA décider à notre place ? Un ex-McKinsey répond

Should we let AI decide for us? A former McKinsey consultant answers

by Renaud Dékode · View the channel on PSI

Humanities, Social Sciences & Thought AI & Society UBDigital and Information technology: general topicsUBJDigital and information technologies: social and ethical aspects
🎙 Renaud Dékode 👥 270K 📅 October 1, 2026 ⏱ 38 min 👁 21 📄 expert opinion 🧭 2026-10-01
Available in: English (current) Français

Keywords

agentic AIautomationupskillingEuropean sovereigntyproductivity

Summary

In this interview, Renaud Dékode welcomes Eric Hazan, former Senior Partner at McKinsey and co-founder of Ardabelle Capital, to discuss whether artificial intelligence should be allowed to make decisions in place of humans. The conversation opens with a historical look at digital transformation in French companies, illustrating how large groups have slowly adopted digital tools over two decades, mainly through heavy multi-year projects and large dedicated teams. Hazan then addresses the arrival of generative AI, noting that while every executive now wants to adopt AI, only a few companies have deployed it at scale. He refutes the idea of a ‘job apocalypse’, arguing instead that AI will disrupt work, automate roughly 30% of working hours over the next decade, and force a massive reorganisation of tasks. Drawing on David Autor’s work, he explains that AI will democratise expertise, enabling less-specialised workers to perform higher-level tasks if properly trained. The discussion then turns to the need for a ‘Marshall plan’ for skills, the generational shift towards lifelong learning, and the risk of a widening gap between large companies and SMEs in France. Hazan also comments on recent announcements from Microsoft and OpenAI about autonomous agents, and outlines two possible corporate strategies: reducing headcount for short-term profitability, or upskilling employees for sustained growth. He advocates the latter, referring to Daron Acemoglu’s concept of ‘pro-worker AI’. Finally, the interview touches on European digital sovereignty, recalling a 2026 episode where Europeans lost access to advanced Anthropic models for 18 days, and the need for a European strategy on energy, data centers and critical infrastructure.

260 words

Critical Evaluation

Value of the Information & Strength of the Argument

The value of this information is high for listeners concerned with the societal and corporate impact of AI. Eric Hazan brings an insider perspective, drawing on his long experience at McKinsey, his participation in large-scale economic studies, and his current venture as an investor. The argumentation is structured and generally sound: he distinguishes between transformation before AI and after AI, separates productivity gains from job destruction, and insists on the strategic choice between short-term cost reduction and long-term upskilling. He relies on credible references such as David Autor, Daron Acemoglu and Philippe Aghion, and explicitly avoids alarmist rhetoric, which strengthens his credibility. The reasoning on the non-existence of a ‘job apocalypse’ is nuanced but open to discussion, as he acknowledges a 15-to-20-year window that could nevertheless be socially painful. The discussion of European sovereignty is relevant and well-argued, even if it relies on an anecdotal event rather than broader evidence.

Scientific Rigor, Source Quality, Title Accuracy

The interview demonstrates a good overall scientific rigour, maintaining a clear distinction between factual observations and opinion. Eric Hazan frequently qualifies his statements, using expressions like ‘roughly’ and ‘about’, which is welcome in such a contested field. The sources mentioned are mainly academic economists and public models, but there is no systematic citation of specific studies or published data; the 27% and 45% automatable-task figures come from Hazan’s own McKinsey studies, yet their methodology is not explained in the video. The title is perfectly aligned with the content, and the central question ‘should we let AI decide?’ is addressed head-on in the conclusion, where Hazan argues that certain decisions must remain human, although he does not elaborate much on that final point. The presence of promotional content for the channel’s forum and community is discreet and does not interfere with the substance of the exchange.

310 words

Title / Content Match

The title accurately reflects the content: the interviewer directly asks Eric Hazan whether AI should be allowed to decide for us, and the entire interview revolves around AI-driven decision-making, its limits and its corporate implications.

Quality & Reliability

8/10

The interview features Eric Hazan, a former McKinsey Senior Partner with 17 years of experience advising large European groups on digital transformation. His economic projections are grounded in known academic work (David Autor, Daron Acemoglu, Philippe Aghion) and his own consultative studies. However, the claims are mostly illustrative and not systematically referenced with verifiable data, and some figures (e.g., 27% of tasks automatable by 2030) appear in the description but are not clearly substantiated in the conversation itself.

Key Moments

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

Cited Sources

  • Renaud Dékode — website — Mentioned in the video description as the official website for discussions and community.
  • Klub Renaud Dékode — Promotional link in the description for the channel's paid community and e-learning platform.
  • Faut-il encore décider ? (book by Éric Hazan & Olivier Sibony, Flammarion) — Referenced in the video description and in the introduction of the interview as the book co-authored by the guest.

Concurring Sources

  • David Autor's research on AI and labour markets — Explicitly cited in the interview to support the idea that AI democratises expertise and reconfigures tasks.
  • Daron Acemoglu's concept of 'pro-worker AI' — Referenced by Éric Hazan as the preferred path for sustained growth through skills transformation.

Contribution & Novelties

The video provides a rare insider view of how French and European corporate boards are actually approaching AI, based on 17 years at McKinsey and a practitioner’s investment perspective. It clearly distinguishes between post-digital transformation fatigue and the specific acceleration brought by generative AI, and maps out two competing corporate strategies: mass headcount reduction versus systematic upskilling, with strong economic arguments favouring the latter. The mention of European sovereignty issues and the 2026 Anthropic access episode adds a concrete and timely dimension that is often absent from such interviews.

Pour aller plus loin :

  • Daron Acemoglu — Wikipedia — Nobel Prize economist whose ‘pro-worker AI’ concept is explicitly referenced in the interview; highly relevant for understanding the debate.
  • David Autor — Wikipedia — MIT economist cited for his research on AI and the democratisation of expertise, central to the idea that AI can augment rather than replace workers.
  • Philippe Aghion — Wikipedia — French economist mentioned for his work on innovation and productivity growth, supporting the argument that AI adoption is necessary for competitiveness.
  • Yann LeCun — Wikipedia — AI pioneer quoted on the importance of ’learning to learn’, directly tied to the interview’s emphasis on lifelong education.
  • Olivier Sibony — Decision-making scholar and co-author of ‘Faut-il encore décider ?’, discussed without a stable URL; his broader work on cognitive biases in boards is complementary reading on this topic.

229 words

Radar Profile

The profile shows a solid quantitative and qualitative information density with a high technical level, although the level of detail remains accessible to a broad audience. Reliability is strong but limited by the absence of original data in the video itself, leaving the viewer to rely on the guest's testimony and references.

Reliability 7/10