The AI jobs panic is colliding with a stubborn reality

Barely a month goes by without another warning that artificial intelligence is about to empty out the world’s office towers, call centers and shop floors. Executives say it in earnings calls. Politicians repeat it at podiums. Surveys now show it sits near the top of the public’s list of anxieties about technology. And the announcements keep coming: this month it was Meta unveiling an autonomous AI system built to run sales, customer service and scheduling, the kind of work that, until recently, only people did.

For all the latest headlines follow our Google News channel online or via the app.

The message is confident and consistent. The machines are coming for the jobs, and soon. What gets far less attention is a simpler, harder to ignore fact: the systems being sold as tireless digital employees keep failing in embarrassingly human ways the moment they are actually put to work.

An airline’s chatbot promised a customer a refund the company never intended to honor, and lost the resulting legal dispute after a court ruled the airline was bound by what its own bot had said. A major fast-food chain ended its AI drive-through ordering test after videos circulated of bot tacking absurd quantities of food onto simple orders. These are not obscure bugs buried in a changelog. They are public, viral illustrations of something more fundamental than a coding error. The large language models underpinning today’s AI boom are not reasoning the way a competent employee reasons. They are built to produce plausible sounding output, not verified, accountable judgment. There is a real difference between a system that sounds confident and one that is right, and much of the current hype quietly glosses over it.

None of this means the technology is a mirage, and it would be a mistake for this region in particular to read it that way. Saudi Arabia has staked genuine political capital on getting this right. The Cabinet designated 2026 the Year of Artificial Intelligence, SDAIA-linked programs have trained more than 11,000 specialists, as the national strategy targets more than 20,000 by 2030, and AI is projected to contribute over $135 billion to the Kingdom’s GDP by 2030. This is not a side bet. It sits alongside the giga-projects and the Public Investment Fund’s diversification push as one of the pillars of Vision 2030 itself.

Which is exactly why the experience of the past year inside the Kingdom’s own ministries deserves more attention than it has received. When SDAIA published its AI Adoption Framework in November 2025, it did not read like a victory lap. It read like an acknowledgment of hard lessons learned. The framework requires public sector entities to build data governance, model accountability, transparency and, crucially, human oversight before AI tools go anywhere near citizens. That requirement did not appear from nowhere. Government data systems across the Kingdom remain fragmented, agencies have historically guarded their own records, and early pilots for citizen facing services such as automated ID renewal and predictive fine payment ran into exactly the kind of friction one would expect when a system trained to sound helpful meets a bureaucracy built on precise, accountable rules. As Arab News reported in late June, officials across the Kingdom now describe the challenge less as a technology gap and more as an execution gap, the harder work of turning promising pilots into systems that function reliably at scale.

That is not a story of failure. It is about maturity and accountability, and it is a more honest one than the breathless version peddled by vendors on both sides of the Atlantic. The ministries that have had real success, the Ministry of Health’s AI supported medical consultations, the Ministry of Justice’s digitized court dashboards, share a common feature. They were deployed into processes that were already well defined, with humans still positioned at the point where judgment and accountability actually matter. The lesson is not that ambition should be scaled back. It is that ambition without governance is how expensive pilots quietly stall.

For Gulf economies racing to build AI driven industries and reskill their workforces, the takeaway is not to slow down but to be precise about where automation genuinely delivers and where it merely performs confidence. Saudi regulators have effectively already reached this conclusion in the frameworks now governing procurement and deployment across government. The rest of the region, and the companies operating within it, would do well to take the same lesson seriously before a chatbot’s promise, made confidently and wrongly, becomes their own courtroom transcript.

The industry leaders promoting fully autonomous AI employees have every commercial incentive to make the leap from impressive demo to inevitable replacement sound smaller than it actually is. Workers, employers and policymakers across the Gulf and the wider Arab world would do well to notice the gap between the sales pitch and what is actually happening on the ground, from a Riyadh ministry office to a fast food drive-through half a world away, and to build their AI strategies around the technology’s real, considerable strengths rather than around the parts of the pitch that keep quietly failing in public.

Read more:

Can Israel do without America?

Escalation against neighbors, Iran’s strategic miscalculation in the Gulf