AI and Your Job in the Gulf: An Honest 2026 Assessment

The numbers, separated from the narrative

Two stories about AI and work are competing in the public discourse. One says nothing real has happened — AI is overhyped, productivity numbers are flat, layoffs are about interest rates and post-pandemic over-hiring, and "AI" is just convenient cover. The other says the great displacement has begun — 60,000 jobs already lost to AI in 2026, white-collar work is the next manufacturing, the recession is here, just not evenly distributed yet.

Both contain a fragment of truth and both are unhelpful for making real decisions about your career.

Here are the figures that hold up to scrutiny.

Employer-disclosed AI-attributed layoffs in 2026 totalled approximately 54,836 across 45+ companies that explicitly cited AI as a driver, with researchers modelling actual AI-displaced or foregone positions at 200,000 to 300,000 across the US economy when including hiring freezes, attrition without backfill, and roles that were never opened (Programs.com AI Layoffs Tracker, 2026). The gap between the two numbers is partly real economic substitution and partly what Built In and others have documented as "AI washing" — companies attributing layoffs to AI when the underlying reason is post-pandemic overhiring, slower revenue, or activist investor pressure, because "AI restructuring" is more palatable to the market than "we hired too many people in 2022" (Built In on AI washing, 2026).

The named numbers from real announcements: Citigroup is reducing headcount by approximately 20,000 in a multi-year overhaul tied to middle- and back-office automation. Accenture announced cuts of around 11,000 roles tied to "how work is changing inside the firm." Baker McKenzie disclosed reductions of 600 to 1,000 employees, up to 10% of its global workforce. McKinsey laid off about 200 internal technology and support staff after automating non-client-facing work. Amazon cut roughly 30,000 corporate roles, though CEO Andy Jassy said publicly these "are not really AI-driven, not right now at least" (Programs.com tracker) (Metaintro analysis, 2026).

The macroeconomic picture is murkier. Goldman Sachs' equity research published in March 2026 found "no meaningful relationship between AI and productivity at the economy-wide level," with the exception of two narrow use cases (call centre automation and software code generation) where productivity gains of around 30% are real (Fortune on Goldman, March 2026). At the same time, Goldman's own surveys show that workers at companies with enterprise AI accounts save 40 to 60 minutes per day, with 75% saying they can now complete tasks that were previously impossible (Fortune on Goldman productivity, April 2026).

The reconciliation: AI is producing real, sometimes large, productivity gains for individuals. Those gains are not yet showing up in aggregate productivity statistics, partly because adoption is still concentrated in a minority of firms (Goldman estimates 80% of companies are not using AI meaningfully yet) and partly because firms are absorbing the gains as headcount reductions and margin expansion rather than output growth.

The honest read: this is a slow-then-fast curve. The first 18 months will look like a normal restructuring cycle with AI as the stated reason for some of it. The next 36 months will start to look structurally different.

What the Anthropic data actually shows

The most rigorous public dataset on AI's labour-market footprint is the Anthropic Economic Index, which uses anonymised conversations on Claude.ai and the Anthropic API to measure what people are actually using AI for at the task and occupation level (Anthropic Economic Index).

The headline finding from the most recent report (which analysed February 2026 usage building on the November 2025 baseline): about 49% of occupations have already had at least a quarter of their tasks performed using Claude in production usage (Anthropic Economic Index, March 2026 report). For a smaller set — including data entry workers, database architects, and parts of computer programming, market research, and financial management — Claude shows proficiency in large swaths of the underlying work.

The pattern is consistent with Anthropic's earlier "Labor market impacts" research: theoretical exposure to AI for white-collar knowledge work is very high, but actual deployment is a fraction of what is technically feasible (Anthropic on labor market impacts). The lag between theoretical and actual matters. It is the gap inside which careers are made, lost, and repositioned.

Two more findings are useful for GCC professionals.

First, the dominant AI use case remains coding: tasks tied to Computer and Mathematical occupations account for roughly 35% of all Claude conversations (Anthropic Economic Index, March 2026). That has implications for the GCC software market, where the practical question is not whether AI affects software work — it does — but how the labour mix between human engineers and AI-assisted productivity shifts.

Second, the speed-up factors are striking. Tasks with prompts requiring only a high-school education are sped up by a factor of 9 in observed Claude.ai usage; tasks requiring a college degree, by a factor of 12 (Anthropic Economic Index, March 2026). These are not 10-20% productivity improvements. These are order-of-magnitude shifts in the time required to produce a unit of work — which is exactly why, for narrow use cases, the displacement is real.

The GCC context: why the local picture is different

It would be a mistake to map the US white-collar layoff narrative one-to-one onto the Gulf. Three structural differences matter.

The first is that the GCC labour market is supply-constrained, not demand-constrained, in most senior professional categories. Saudi Arabia in particular needs more — not fewer — qualified bankers, lawyers, healthcare administrators, and infrastructure engineers to deliver Vision 2030 mega-projects. Saudisation and Bahrainisation policies are pushing employers to hire more nationals, not to cut total professional headcount. Even where AI compresses individual productivity requirements, the absolute demand for trained professionals across the bloc is rising.

The second is that the AI build-out is happening in the GCC, not just being absorbed by it. We covered the G42 / Stargate UAE / HUMAIN / DIFC AI-native financial centre programme in our companion piece. The GCC is one of the largest greenfield AI buyers and builders in the world. That creates direct hiring demand — for AI engineers, data centre operators, AI governance lawyers, model risk professionals — that Western markets are not generating at the same scale.

The third is that public sector employment plays a stabilising role that does not exist in the US. National oil companies, sovereign wealth funds, central banks, regulators, and government services employ a large share of senior professionals in the bloc. These employers adopt AI more slowly and use it more cautiously. Their headcount decisions are driven by mandates and political considerations, not quarterly margins.

The result is that the GCC is not facing the same near-term displacement risk as US white-collar workers, but the medium-term reshaping is real. The roles that exist in 2030 will look different from the roles that exist today. The professionals who navigate the transition successfully will be the ones who reposition early.

Sector-by-sector: where the squeeze is real and where it isn't

Investment banking. The most-discussed change in IB is the analyst experience. Goldman Sachs' internal "Banker Copilot" is reported to cut M&A deal preparation time by approximately 40%, primarily by automating pitchbook drafting, comparable company analysis, and first-pass financial modelling (coverage of Goldman's AI banking strategy). For analyst-class hiring in DIFC and Riyadh, the implication is not "no analysts" — banks still need a pipeline — but "smaller analyst classes per VP and MD." Expect 2026-2028 graduate intakes at the regional bulge brackets to be 15-25% smaller than 2022-2024 vintages, with higher quality bars on entry. At the senior level (Director, MD) the displacement risk is much lower because client coverage, transaction judgement, and relationship management remain human-only.

Management consulting. The single biggest exposure in the senior professional categories. The associate and senior associate tier — the people who do data gathering, slide drafting, and analytical synthesis — overlap heavily with capabilities that Claude, ChatGPT, and Gemini perform well. McKinsey's 200-person internal layoff was an early signal (Programs.com), and Accenture's 11,000-role restructuring is a larger one. Strategy firms in the region (the Middle East practices of MBB and Big Four Strategy) are quietly redesigning their consultant-to-partner ratios. The defensive positioning is to build distinctive sector or functional expertise that cannot be replicated by an LLM with public-domain knowledge — defence and security, sovereign advisory, energy transition, family business governance. The exposed positioning is generalist work where the deliverable is a 60-slide deck synthesising public information.

Legal. Bifurcated. Document review, due diligence, contract first-drafting, and standard advisory memos are being automated rapidly. Magic Circle and US firms in DIFC and ADGM have all rolled out AI platforms (Harvey, CoCounsel, internal builds) and the productivity claims are credible. Baker McKenzie's 10% global cuts are a leading indicator (Programs.com). The protected segments in the region are: contentious work (litigation and arbitration), regulatory and DFSA / DFM advisory, sanctions, capital markets execution, transactional partnership work where human judgement and client relationships dominate. Junior associate hiring at international firms in the GCC is being scrutinised more carefully, but the structural undersupply of qualified senior dispute resolution and regulatory lawyers in the region keeps senior demand strong.

Banking middle and back office. This is where displacement is happening at scale, globally and in the GCC. Citigroup's 20,000-headcount reduction is the canonical example, with a significant share attributed to operations, reconciliations, KYC, and routine risk reviews (Programs.com). GCC banks — Emirates NBD, FAB, ADCB, Saudi National Bank, QNB — are running similar programmes more quietly. Operations roles below VP level, particularly in standardised functions, are the highest-displacement category in the regional banking sector. The defensive move is to shift toward client-facing relationship banking, complex product structuring, financial crime investigations (which still require human judgement on novel cases), or to move into the AI-adoption side of the business itself (model risk, validation, AI governance).

Audit and accounting. Big Four practices in the GCC are restructuring. Audit work — sample selection, substantive testing, draft reporting — is increasingly AI-assisted. Headcount at the associate and senior associate level is under pressure across the region, even as demand for senior audit partners (who sign opinions and own client relationships) remains stable. Tax and advisory work in specialised areas (transfer pricing, VAT, corporate tax, ESG assurance) is more insulated.

Marketing and communications. Content production, ad copy, social media management, basic graphic design, and first-draft strategy work are all being absorbed by AI tools at GCC agencies and in-house teams. Junior creative and content roles are the most exposed in the bloc. Brand strategy at the senior level, executive communications, regulatory and crisis comms, and physical experience design (events, retail) are more defensible.

Tech and engineering. Counter-intuitive: senior software engineers, AI engineers, ML practitioners, and data scientists are seeing demand increase in the GCC, not decrease. The Stargate UAE campus, HUMAIN's data centre programme, and the DIFC AI-native build-out are all hiring aggressively. The squeeze in tech is concentrated at junior coder level — companies that previously hired large junior engineering classes to do routine work are now hiring smaller, more senior teams that use AI to multiply output. If you are a senior engineer with AI engineering depth, the GCC is the strongest market in the world for you right now.

Healthcare administration, supply chain, real estate, energy operations. Less affected at the senior level. These sectors involve significant physical, regulatory, and relationship components that AI does not yet handle. Junior analytical roles are exposed; senior operational and decision-making roles are not.

What to actually do if you are senior in an exposed seat

The professionals who navigate this transition successfully will not be the ones who use ChatGPT a bit more or take an AI literacy course on LinkedIn. The structural moves that matter are these.

Move up the judgement curve, not just up the title. Promotion alone does not protect you. What protects you is doing work that requires judgement informed by experience that LLMs cannot replicate — client relationships, novel transactions, regulatory advocacy, deal structuring, people management at scale. If your senior role mostly involves reviewing analyst output and approving decks, you are exposed. If your senior role involves making non-obvious calls under uncertainty with imperfect information, you are not.

Build deep sector expertise that compounds. Generalist white-collar work is where AI hits hardest. Domain depth — Saudi banking regulation, ADGM digital assets, GCC family office structures, regional energy transition, healthcare regulation — is harder to displace because the training data is shallow and the consequences of error are high.

Get inside the AI adoption side of your industry. Every bank, law firm, consultancy, and corporate in the GCC is building or buying AI. The roles at the intersection — heads of AI for legal, AI governance for banking, model risk validation, AI ethics, AI procurement — are the fastest-growing professional categories in the region. These seats did not exist in meaningful numbers two years ago. They are being filled now.

Use AI heavily yourself, even if your firm is slow. The fastest way to understand where displacement happens is to do your own job with AI assistance for 90 days and notice what you stop needing to do manually. That observation is more valuable than any consultant's report on AI strategy.

Maintain your network across firms and geographies. If your firm restructures, the cost of finding the next seat is dominated by the strength of your professional network. Senior professionals who lose roles in restructuring cycles and land within 90 days are almost always the ones with deep regional networks. The ones who take 12+ months to find a comparable seat are usually the ones who outsourced their career management to their employer for a decade.

Be honest with yourself about what you do all day. This is the hardest one. Make a list of the top 10 tasks that consume your week. Score each one for how much of it Claude, ChatGPT, or Gemini could plausibly do today, and how much in 18 months. If more than half your week is in the "AI can plausibly do this" column, you have a repositioning project.

What this means for the next 18 months

The defining feature of AI's impact on Gulf white-collar work over the next year and a half is that the headline displacement numbers will continue to come from the West, while the structural reshaping will continue to happen in the GCC at a slower but more permanent pace. Citigroup's 20,000-person reduction is a globally relevant data point (Programs.com), but the more important signal for someone in DIFC or Riyadh is the quiet redesign of analyst classes, the slimming of associate cohorts at law firms, and the merger of teams in middle-office banking that is already underway across regional employers.

Fortune's reporting on Anthropic's labour market research summarised the situation as well as anyone has: "AI can already do a huge portion of many jobs," but the actual labour-market translation will play out over years, not months (Fortune on Anthropic, April 2026). The professionals who use those years to reposition will end up in roles that pay more and are more secure than they have today. The professionals who wait will not.

The GCC's particular advantage is that the regional economy is in expansion mode, not contraction mode. Demand for senior professional talent across finance, energy, technology, infrastructure, healthcare, and government is rising structurally, even as the specific composition of the work shifts. The question is not whether you will have a career in the Gulf in 2030. The question is whether the version of your career that exists then is one you positioned yourself for, or one that happened to you.