Per la versione italiana di questa lettera, clicca qui.
Dear Mr Gates,
I read your essay “The turbulent AI era is here. The choices we make now are critical” with more agreement than I might once have expected.
I share several of your concerns. Artificial intelligence may destroy many jobs permanently; it may concentrate power where power is already concentrated; it may make fraud, surveillance, manipulation and violence easier; it may weaken critical thinking and interfere with the difficult human work through which children and adults learn to relate to one another. I also appreciate your acknowledgement that you have benefited enormously from the technology industry, and your insistence that decisions about our common future should not simply be left to AI companies.
I do not agree with everything you write. But I believe the most important disagreement lies even deeper than the policies you propose. Before asking how we should manage the transition to AI, I think we need to ask two prior questions: What is this technology turning us into? And what kind of economic and political system has produced it?
Technology is never neutral
Marshall McLuhan's famous phrase, “the medium is the message,” points toward something that is too often forgotten in discussions about artificial intelligence. A technology is not merely a neutral instrument waiting for a good or bad human intention. Its form changes the environment in which we live, the habits we develop, the things we consider normal and, eventually, ourselves.
You write that AI can now “replace and even exceed human cognition.” I think we should stay with those words for a moment.
Thinking is not merely the production of a correct answer. It is also the slow and sometimes painful process through which a person encounters uncertainty, makes mistakes, develops judgment, discovers limits, learns patience, takes responsibility and gradually becomes aware of who he or she is. Writing is not merely the production of text. Remembering is not merely retrieving information. Conversation is not merely exchanging useful sentences. These activities participate in the formation of a human being.
If we increasingly delegate them to machines, we should not assume that only the output changes while the human being remains untouched.
Whatever the intentions of individual designers may be, the most celebrated capability of today's AI is precisely its ability to take over cognitive work that previously required a person. And the prevailing economic incentive is to reduce the number of steps, shorten processing times and remove operational obstacles, lower labour costs, accelerate production and make human intervention unnecessary wherever possible.
This creates a danger that cannot be solved merely by making AI safer or more equitable. A tool designed and rewarded for replacing cognitive effort may gradually weaken the very faculties through which human beings grow. You yourself use the valuable expression “productive struggle” when discussing education. I would extend that insight far beyond the classroom. Much of what makes us human is born through productive struggle: the struggle to understand, to remember, to express ourselves, to tolerate frustration, to meet another person who does not behave as we wish, and to remain present when no machine supplies an immediate answer.
So the question is not only: What can AI do for us? It is also: What does habitual dependence on AI do to us?
AI did not emerge outside our economic system
My second concern is that AI is often discussed as though it were an external force suddenly arriving and disrupting an otherwise neutral society. I see it differently. Artificial intelligence is a product of the economic and political order that created it.
Our present system rewards competition, accumulation, endless expansion, concentration of ownership, reduction of labour costs and the conversion of ever more areas of life into markets. It encourages companies to replace people when replacement is profitable. It rewards those who accumulate enough capital to acquire still more capital, infrastructure, data and political influence. Environmental and human costs can easily become externalities carried by someone else.
In this sense, AI does not create the underlying logic. It accelerates it.
You describe a “vicious cycle” in which one company adopts AI or robots, lowers its costs, and forces competitors to do the same. You also recognize that, without intervention, the benefits will accrue to a small group. I agree. But this leads me to a more radical conclusion: we cannot adequately mitigate the consequences of AI without reconsidering the foundations of the economic system that makes this vicious cycle rational.
I believe we need to move away from an organized war of each against all and toward a genuinely cooperative social order, one in which economic activity is subordinated to the dignity and freedom of every person rather than the reverse.
That also requires us to face extreme differences in wealth and income. A society cannot credibly speak of equal freedom while one person possesses resources and influence comparable to those of institutions and millions of other people possess almost none.
I say this to you with respect, not as a personal insult. Your own extraordinary wealth makes the question concrete rather than theoretical. Through the fortune you accumulated and the foundation you created, you have acquired a capacity to influence research, public health, technology and political debate that very few human beings possess.
I do not doubt that a person may try to use such power for good. But benevolent concentration of power is still concentration of power. A system should not have to depend on the wisdom or virtue of a few extraordinarily wealthy individuals in order to serve humanity.
Beyond “Human Reserved”
One of the most interesting proposals in your essay is what you call Human Reserved: choosing to preserve certain activities for human beings even when machines could perform them.
I would like to take that idea further.
What if we reserved for human beings not only some jobs, but something more fundamental: dignity, conscience and freedom?
There should be parts of human existence that no government, corporation, algorithm, expert body, employer, philanthropist or majority is entitled to appropriate. The inner work of conscience should be Human Reserved. The right to think, doubt, refuse and change one's mind should be Human Reserved. The body should be Human Reserved. The possibility of living according to one's deepest convictions, so long as one does not deliberately exercise violence against others, should not depend on wealth, status or obedience.
If we are prepared to accept a little economic inefficiency in order to protect human work, as you suggest, should we not be prepared to accept much more in order to protect human freedom?
Medicine, vaccines and the boundary of conscience
This brings me to an issue that has long been central to your foundation's work: vaccines and public health.
My position begins before any dispute over the merits of a particular vaccine. It is an ethical principle: no medical intervention should be imposed on a human being against his or her conscience. The same principle applies when parents are placed under coercion regarding medical or preventive interventions for their children.
To compel another human being to violate his or her conscience is itself a profound form of violence, even when the coercion is exercised in the name of a good end.
For me, non-violence cannot be a principle that applies only when it is convenient. If we make exceptions whenever we are sufficiently convinced that our goal is beneficial, then the principle gradually disappears. The person using coercion will almost always be able to describe the intended end as necessary, protective or benevolent.
Science does not resolve this moral question for us. Science can investigate effects, probabilities, mechanisms, benefits and risks. It can produce evidence of varying strength. But it cannot answer the prior ethical question: Who ultimately has authority over a person's body and conscience?
There is another reason for humility. Scientific knowledge is not a collection of final truths. Its dignity lies precisely in its capacity to doubt, test, replicate, correct and sometimes overturn what was previously accepted. In my recent article When Science Stops Doubting: Falsehoods, Fraud and the Crisis of Contemporary Research, and in a recent interview, I discussed problems of reproducibility, distorted incentives, conflicts of interest, publication bias and the danger of replacing scientific method with authority.
The conclusion is not that science should be rejected. Quite the opposite. It is that no institution should ask people to “believe in science” as though science were a creed. We should ask: What are the data? How were they obtained? What are the uncertainties? What findings contradict them? Who funded the research? Can independent people reproduce the result? What would show us that we are wrong?
And even when the evidence for a medical intervention is strong, the strength of an empirical argument and the moral right to compel another person are two different questions.
Freedom must not be a privilege of the rich
Defending freedom of medical choice does not mean abandoning people to themselves. Quite the opposite.
A freedom that only wealthy people can exercise is not much of a freedom.
I would like to see health systems that offer everyone, regardless of income, genuine access to a plurality of medical and preventive approaches; systems that give serious attention to primary prevention, nutrition, movement, psychological and social wellbeing, and non-pharmaceutical as well as pharmaceutical approaches; systems that make evidence, uncertainty, risks and alternatives transparent; and systems in which declining one intervention does not mean losing the right to care, dignity or social participation.
The task of a humane health system should be to inform honestly, to make meaningful choices materially accessible, to care for those who suffer, and then to respect the decision of the person.
This is where your concern for global equity could become even more ambitious. Rather than merely asking how we can make particular technologies, drugs or vaccines reach more people, we could ask how every person can gain access to care without surrendering conscience.
That would be another kind of Human Reserved: not a protected profession, but protected human sovereignty.
Using power to disperse power
Near the end of your essay, you explain that you intend to use your voice, your time and the Gates Foundation's resources to put AI and equity higher on the political agenda.
I would respectfully invite you to consider an additional possibility: use extraordinary power not merely to do good with power, but to create a world in which extraordinary concentrations of power become less necessary and less possible.
Support institutions that disperse decision-making rather than centralize it. Support scientific structures in which inconvenient questions can be investigated without researchers risking their livelihoods. Support health systems that give poor people real choices rather than a single officially favoured path. Support economic arrangements in which the difference between the poorest and the richest is not so vast that wealth itself becomes political authority. Support technologies that strengthen human capacities rather than making human beings progressively dependent on systems they neither own nor understand.
I do not believe lasting peace is compatible with a world permanently divided between those who can shape the conditions of other people's lives and those who must simply live under conditions shaped for them. As long as extreme material inequality persists, so will the structures that generate humiliation, domination, conflict and war.
This does not mean that everyone must be identical or possess exactly the same things. It means that no person's economic power should become so great that another person's freedom becomes small by comparison.
A question for both of us
I am not writing this letter in order to defeat you in an argument. I do not claim to possess final truth. In fact, one of the principles I try to apply to myself is that I should be among the first people whose ideas I question.
If I am wrong, I hope reality will correct me. If you are wrong, I hope you remain free to discover it. And I want that same freedom for everyone: the freedom to doubt, to investigate, to refuse, to change one's mind and to seek a different path without being economically or socially destroyed for doing so.
I have written elsewhere about non-violence without exceptions. For me, non-violence is not passivity. It is a refusal to build peace through domination. It asks us to examine not only the result we want, but the intention and the means by which we seek it.
This is why I think your final question—how we preserve our humanity through the AI transition—is even larger than AI.
We preserve our humanity when we refuse to treat another person's conscience as an obstacle to be overcome. We preserve it when economic efficiency does not outrank human dignity. We preserve it when knowledge remains open to doubt. We preserve it when technology serves the development of human capacities instead of quietly replacing them. And we preserve it when those who possess great power are willing to question the structures that gave them that power.
Perhaps, then, the most important Human Reserved domain is not a category of employment.
Perhaps the human person must be Human Reserved.
Dear Mr Gates, we may disagree profoundly on some matters. Nevertheless, I sincerely wish you well, just as I wish well to those who agree with me and those who do not. Precisely because your influence is so great, I hope you will consider these questions not as an attack, but as an invitation.
You ask how AI can leave humanity more equal and more human.
My answer is that we must begin before AI: with the economic order that produced it, with the concentration of power that directs it, and with an unconditional respect for the dignity, conscience and freedom of every person.
With respect and hope,
Francesco Galgani,
August 30, 2026