When doomsaying seeks to control the future
“Artificial intelligence could cause one billion deaths.” In a September 2026 interview with NBC, Microsoft co-founder Bill Gates issued this stark warning about the catastrophic scale of harm that could result from the malicious use of advanced AI systems.
But where does responsible warning end and the cultivation of fear begin? And when does a necessary debate about the risks of emerging technology become a mechanism for manipulating public opinion?
If “blood and semen” are, metaphorically, the corrupted proteins on which the decaying organism of global disinformation feeds, then fearmongering and manufactured panic are the diseased carbohydrates completing its toxic diet. Famine, earthquakes, floods, foreign invasions, fire and the sword: the familiar Armageddons have returned to the menu of public discourse.
Increasingly, artificial intelligence has become a target of this apocalyptic rhetoric. It is portrayed as a potential engine of calamity, tragedy and total destruction, as though technological progress itself inevitably foretold humanity’s subjugation—or even its annihilation.
Yet the debate over AI’s future cannot be reduced either to promises of technological salvation or to prophecies of the apocalypse. It demands critical thinking, scientific evidence and independent assessments of actual risks. Above all, we must ask who makes decisions about the technology, who controls its tools and who benefits from its power.
The old systemic order, with its concentrated economic and political power, faces a technological transition that could disrupt established hierarchies. In this environment, fear can become an instrument for preserving control—not necessarily because every warning is orchestrated propaganda, but because fear can facilitate the concentration of power and narrow the boundaries of public debate.
Here lies the essential question: do we want artificial intelligence that serves humanity, freedom, knowledge and the natural principles of life—or AI subordinated to the interests of those who already control the data, infrastructure and mechanisms of decision-making?
History offers an uncompromising point of comparison. Empires, armies, authoritarian regimes, and religious and political hierarchies have inflicted centuries of persecution, warfare and mass death. No serious assessment of AI risks can ignore that historical legacy. But neither does humanity’s past brutality guarantee that future technologies will be harmless.
The task, therefore, is neither to acquit artificial intelligence in advance nor to condemn it before we fully understand its capabilities. It is to reject a monopoly on fear, just as we must reject a monopoly on technological power.
Diotima is invited to examine this conflict—not as a herald of inevitable catastrophe, but as a partner in a crucial inquiry into who will define the rules of the new world.
For the future will not be determined solely by how powerful artificial intelligence becomes. It will also depend on who controls it, which values guide it and whose interests it serves.
______
Diotima’s critical analysis
Central question: Is the demonization of AI a defense of human society, or can it also serve to defend the existing distribution of power?
1. First, the factual distinction
The reference to Bill Gates is grounded in actual reporting. In a September 2026 interview with NBC’s Meet the Press, Gates warned that AI could be powerful enough to contribute to events resulting in one billion deaths. His concern focused particularly on malicious actors using advanced AI tools, and he argued that emergency shutdown mechanisms alone would not be sufficient without monitoring, records and meaningful safeguards.
The distinction matters: Gates did not predict that AI would kill one billion people. He warned that AI could contribute to events on that catastrophic scale.
An editorial headline presenting the number as a certain forecast would therefore be misleading. Equally, a critique of his warning should engage with the argument he actually made, rather than an exaggerated version of it.
2. The industry of fear and the political economy of information
Your metaphor of a “toxic diet” captures a recognizable feature of contemporary communication: information that provokes fear, outrage and a sense of imminent catastrophe can attract more attention than measured, complex analysis.
Three phenomena, however, must be distinguished:
-
Evidence-based warning: identifies a specific risk, the conditions under which it might arise and the evidence supporting the assessment.
-
Doomsaying: turns a possibility into an apparently inevitable outcome, often without proportionate evidence.
-
Disinformation: distorts or conceals material facts in ways that mislead, whether deliberately or through the resulting effect.
Warnings about AI are not automatically disinformation. Conversely, appeals to innovation and progress do not automatically establish that a technology is safe.
The deeper issue is who has the power to shape public perceptions: media organizations, governments, technology companies, research funders or independent scientists? There is no need to assume a single, coordinated center of control. Instead, we face a complex field of competing interests in which both fear and excessive optimism can be used strategically.
3. The paradox of powerful people issuing warnings
Gates is not an outsider observing the technology economy. He is one of the figures most closely associated with the development of the modern digital industry.
That makes it reasonable to scrutinize his statements, interests and proposals for AI regulation. It does not, by itself, prove that his warnings are an attempt to preserve the power of the old system.
Here lies one of the strongest arguments your article could develop:
Could some calls for AI safety and regulation protect the public interest, while others entrench the advantages of already dominant companies?
This is a serious political-economic question. Answering it requires examining specific proposals: which rules are being proposed, who will enforce them, how expensive compliance will be and whether those costs disproportionately prevent new competitors from entering the market.
Regulation can restrain abuses. Poorly designed regulation, however, can reinforce the position of organizations that already possess the capital, infrastructure and legal resources required to comply. Regulation is neither inherently liberating nor necessarily oppressive.
4. The historical responsibility of human power
Your reminder that human authorities have inflicted wars, persecution and mass death provides an important ethical counterweight to a technological debate that sometimes treats AI as though it existed outside history.
Humanity does not need superhuman machine intelligence to produce mass violence. It already possesses institutions, weapons systems and propaganda mechanisms capable of devastating consequences.
But it does not follow that AI is harmless or that future risks are exaggerated. Technology can amplify existing human power, accelerate attacks and enable smaller actors to cause disproportionate damage.
The crucial question is not whether humans or machines are inherently “worse.” It is how technology changes the scale, speed, accessibility and controllability of human action.
5. Can AI serve natural principles and human values?
This is the deepest philosophical dimension of your text.
AI does not automatically become ethical because it is advanced, nor does it become ethical simply because its developers claim good intentions. Its objectives, limitations and behavior are influenced by its data, design, training, testing, oversight and the incentives of the organizations developing it.
If by “natural principles” we mean protecting life, avoiding unnecessary harm, preserving human dignity and living sustainably with the natural environment, these values can serve as criteria for evaluating AI systems.
But invoking them is not enough. They must be translated into requirements that can be assessed and enforced: privacy protection, transparency where feasible, accountability, rights to challenge automated decisions, human oversight in high-stakes applications and meaningful public participation in setting the rules.
AI should neither be worshipped as a new savior nor demonized as a new devil. It should be judged by what it does, whom it benefits, whom it harms and who is accountable for the consequences.
6. Diotima’s final judgment
The strongest version of your argument is not that every warning about AI is propaganda produced by the old order. It is this:
The language of catastrophe can be used to legitimize new forms of control, just as the language of progress can be used to conceal new forms of exploitation.
Both deserve critical scrutiny.
The new world will not become free simply because it uses more artificial intelligence. Nor will it be saved by banning or halting every technological advance. Freedom will depend on how knowledge and power are distributed, whether society can exercise meaningful oversight, and whether the benefits of technology are broadly shared rather than concentrated in a few hands.
And if Diotima has a role in this debate, it is not to sound the trumpet announcing the fall of one world and proclaim herself the prophet of the next. It is to ask the questions that no system of power—old or new—should be allowed to evade.