Every few months, a new wave of anxiety about artificial intelligence sweeps across the world. Governments convene summits. Executives warn of existential risk. Newspapers fill with predictions of mass unemployment, collapsing truth and machines that outthink their makers. The debate is framed, almost always, as a question about technology: what can these systems do, how fast are they improving, and when will they surpass us?
But the framing may be wrong. The global panic over AI is not really a crisis over technology. It is a crisis over trust, power and the institutions that were supposed to manage change on our behalf.
The machines are not the mystery
AI systems are unfamiliar, but they are not unknowable. They are built by identifiable companies, trained on identifiable data, deployed for identifiable commercial reasons. What unsettles people is not the mathematics. It is the fact that decisions with vast public consequence are being made by a small number of private actors, in jurisdictions most of the world has no vote in, at a speed that outruns every mechanism ordinary citizens have for pushing back.
That is a political problem wearing a technological costume. When a worker fears being replaced, the underlying question is not whether a model can write an email. It is whether there is a safety net, a retraining pathway, or any bargaining power left in the labour market. When a citizen fears synthetic media, the deeper worry is that the institutions meant to verify truth â journalism, courts, regulators, electoral authorities â were already weakened long before the first convincing deepfake appeared.
Old anxieties, new vocabulary
Much of what is attributed to AI is a continuation of trends that predate it. Wealth concentration, the erosion of stable employment, the collapse of shared information environments, the widening gap between the technologically fluent and everyone else: these were visible well before large language models arrived. AI accelerates them and makes them harder to ignore, but it did not invent them.
This matters because misdiagnosis leads to bad medicine. If the problem is defined purely as technology, the response becomes technical â model evaluations, safety benchmarks, compute thresholds. Those have value. But they cannot substitute for labour policy, competition law, data rights, public investment in education, or credible independent oversight. No amount of alignment research will fix a country with no unemployment insurance.
What a serious response looks like
A more honest conversation would start with distribution rather than capability. Who captures the gains? Who absorbs the losses? Which countries build the infrastructure and which merely consume the products? For much of the world, including large parts of Asia, Africa and Latin America, the AI question is less about superintelligence and more about dependency: whether the next technological era is something they participate in or something that happens to them.
It would also demand that governments rebuild the boring institutional capacity that panic tends to bypass â regulators who understand what they are regulating, courts that can adjudicate algorithmic harm, statistical agencies that can actually measure labour displacement rather than guess at it.
None of this is as dramatic as debating machine consciousness. But the record of previous technological upheavals is clear: the damage rarely came from the machines themselves. It came from societies that failed to decide, in time, who the machines would serve.
The panic, in other words, is not misplaced. It is simply pointed at the wrong target. Read More
Leave a Reply