World must carefully consider AI's purpose - Arab News

TEXT ANALYSIS PROTOCOL

TEXT START: Artificial intelligence is reshaping how we communicate, access information and work, as well as how income and status are distributed and even how we wage war.

THE DISSECTION

Daron Acemoglu, MIT Nobel laureate, uses Pope Leo's encyclical as a springboard to advocate for "human-complementary" AI design — AI that augments rather than replaces humans. The essay argues: (1) current AI trajectory is a choice, not fate; (2) human-AI collaboration outperforms AI mimicking humans; (3) automation-first AI has historically concentrated gains at the top; (4) institutional interventions — antitrust, regulation, public investment — can steer AI toward broad-based welfare. He positions himself as the serious, evidence-based voice cutting through techno-optimism and techno-pessimism alike.

The structure is impeccable. The conclusion is furniture.

THE CORE FALLACY

Acemoglu's central error: treating an institutional failure as a policy problem still solvable by institutions.

He writes as if antitrust action against dominant platforms, public investment in human-complementary AI, and meaningful data rights are live options. He does not engage with the mechanism by which those platforms became dominant — a mechanism that systematically captured the institutions designed to constrain them. Antitrust enforcement has been theoretically available for thirty years. It did nothing. Public investment in alternative AI paradigms is structurally crowded out by private capital with direct access to regulatory bodies. Worker data rights collapse the moment they conflict with shareholder value.

The historical evidence he cites actually destroys his argument. He acknowledges that "four decades of digital automation have already concentrated gains at the top, hollowed out middle-skill work and produced disappointing aggregate productivity growth." He then concludes we should expect differently this time — from an even more concentrated industry. This is not analysis. This is ritual.

The fundamental misdiagnosis: Acemoglu frames the problem as bad design philosophy — the AI industry chose mimicry/automation over augmentation. But the automation-first trajectory is not a philosophy. It is the mathematical output of capital markets optimizing for labor cost reduction at scale. You cannot shame this outcome into changing by appealing to human dignity or offering a more productive alternative. The alternative is less profitable in the short run, which is the only run that matters for capital allocation.

HIDDEN ASSUMPTIONS

1. Institutional voluntarism: Political systems can meaningfully redirect technological trajectories when the primary beneficiaries of current trajectories hold structural power over those systems. Unstated. Fatal.

2. Historical continuity of reform capacity: The assumption that because past rounds of automation were partially managed (with concentrated gains as the result), this round is similarly manageable. Ignores that the scale of cognitive automation collapses the negotiating position of labor entirely, not incrementally.

3. Productivity as the relevant metric: Acemoglu treats "disappointing aggregate productivity growth" from digital automation as a failure. Under DT logic, the concentration of gains IS the outcome. Productivity gains flowing to capital are not a bug — they are the designed result.

4. Labor's structural relevance: The essay implies workers retain enough market power to force human-complementary design through wage pressure, bargaining, or political organizing. No mechanism for this is provided.

5. Cross-class interest alignment: The suggestion that "we can and must demand a different design" implies a shared interest between workers globally and the concentrated AI industry. There is no such shared interest. The interests are directly opposed.

SOCIAL FUNCTION

Prestige signaling from a secure position. Acemoglu is a Nobel laureate at MIT. His essay performs intellectual seriousness — citing Pope Leo, using the "technology is a choice" framing, offering concrete policy prescriptions — while the actual content is harmless to the power structure he nominally critiques. The essay will be read by other academics, policy wonks, and thoughtful professionals who will feel calibrated. Nothing will change. No one at OpenAI, Anthropic, or the relevant regulatory capture nodes will read this and adjust behavior.

The essay's function is to occupy the "serious critique" slot in the intellectual marketplace, thereby preempting more structurally accurate diagnoses that might actually generate political mobilization. It performs dissent without threatening the underlying arrangement.

THE VERDICT

Acemoglu is diagnosing a terminal patient and prescribing rest and fluids. The DT framework says: the patient is dead. The automation-first trajectory is not a design error correctable by institutional intervention — it is the mathematically optimal output of capital markets operating under the structural conditions that actually exist, not the ones that should exist. The institutions Acemoglu calls upon have been captured, defunded, or rendered toothless over exactly the period during which the current crisis developed. Expecting them to reverse the process is not optimism. It is denial with a prestigious veneer.

The essay's most revealing line is the invocation of Phil Jackson's Chicago Bulls — a team built on complementary specialization. Acemoglu uses this metaphor to argue for human-AI complementarity. But the Bulls' success required Jackson's authority to make the players submit to the system. Who has the authority to make AI companies submit to a complementary design mandate? What is the mechanism? The essay never asks. Because asking would mean acknowledging that no such authority exists at the scale required, which would collapse the entire "technology is a choice" frame into "technology is a choice for those who control it, and they have chosen automation."

The final line — "the rest of us must stand up for humanity" — is the tell. "Humanity" as a political actor does not have standing in the decisions being made. Acemoglu knows this. The essay performs awareness of the problem while declining to name the structural cause: the people who control AI have selected automation because it maximizes their power and profit, and no institutional check with actual enforcement power currently exists or is being built.

This essay is ideological anesthetic. It is well-written, intellectually serious, and completely useless as a guide to the actual trajectory. It keeps the conversation in the zone of reformable choices while the relevant choices are being made by a handful of people in a handful of companies with no effective opposition.

FINAL ASSESSMENT

The essay functions as a calibration object for comfortable elites — people who can read "AI is a threat to human dignity," nod along, feel appropriately concerned, and then go back to their positions in a system that will not change in any way the essay recommends. Acemoglu benefits from this arrangement. So do his readers. The AI industry benefits most of all: it gets to be "criticized" by people with no mechanism to enforce the criticism, which makes the criticism look like genuine oversight while the trajectory continues unbroken.

The essay's recommendations — antitrust, public investment, worker data rights — are not wrong as abstract policy. They are irrelevant as predictions of what will happen. DT says: they will not be implemented at the scale required. The lag is short now. The window for institutional correction is closing while Acemoglu writes about Phil Jackson.