Кто управляет ИИ-бумом: экономист Джейсон Фурман о регулировании, рынке и морали

Economist Jason Furman, Harvard professor and former White House economic adviser, gave an interview to Bloomberg on August 2, 2026, calling for an end to a one-size-fits-all approach to AI regulation. According to him, different risks — from biological weapons to job losses — require fundamentally different tools: government prohibition, market adaptation, or broad public moral consensus.

Where Should the State Intervene in AI?

Biological weapons are the primary area requiring mandatory state control, Furman argues. AI systems capable of accelerating the synthesis of dangerous pathogens or designing new disease-causing compounds create an asymmetric risk: a small group of people with access to powerful models can potentially cause damage on the scale of a national catastrophe. The market cannot handle this — commercial players have no incentive to voluntarily limit the dangerous capabilities of their products.

Key points from Furman's Bloomberg interview:

• Biological weapons and AI — an unconditional zone of strict state regulation
• Job losses due to automation — a task for markets and social programs, not for prohibitions
• Not every AI problem needs to be solved with a single law or a single agency
• The main debate about AI will ultimately prove to be moral, not technical or economic

Where Markets Do Better Than Laws

Job losses due to automation are a historically recurring challenge that the economy has overcome before. The steam engine, electrification, and the internet ultimately created more jobs than they destroyed, although the transition period is always painful for specific people and industries. Direct bans on automation, in Furman's view, are counterproductive: they slow productivity growth without saving those who have already lost their jobs.

The state is appropriate here as a shock absorber: retraining systems, expanded unemployment insurance programs, education reform for new professions. But the adaptation mechanism should be shaped by market signals — resources for retraining are directed where there is a labor shortage, not through administrative bans.

This approach assumes that the government does not pick winners among technologies, but provides the conditions under which society copes with change without catastrophic social upheaval.

Why the AI Debate Has Become Moral

The technical and economic questions of AI regulation, in Furman's view, are solvable — though complex. But behind them lies a deeper level: who and on what basis makes decisions that change the lives of millions of people?

According to

Furman, as stated in his Bloomberg interview, "the greatest debate about AI will ultimately prove to be moral" — precisely because it touches on fundamental questions about power, justice, and human autonomy.

How acceptable is mass surveillance for the sake of public safety? Can an algorithm be delegated decisions about loans, hiring, or medical diagnoses? Who bears responsibility if AI makes a mistake? These questions have no technically "correct" answer — they require political and cultural consensus that has not yet formed in any country. This is precisely where AI regulation is fundamentally more complex than nuclear energy or pharmaceutical regulation: there, society more or less knows what outcomes it wants. With AI, that consensus does not exist.

What This Means

Furman's position offers a way out of the "regulate or not" deadlock: the question is not whether the state intervenes, but where exactly, with what tool, and toward what goal. For developers and policymakers, this means one thing: a single universal AI law most likely will not exist — and perhaps should not. Instead — sectoral rules, differentiated by risk level and type of application.