When Will AGI Arrive — And Who Loses Their Job First?

Let's get the uncomfortable part out of the way: we're already living through the early stages of the biggest labour market disruption since industrialisation. Not because robots are welding car doors — that's been happening for decades — but because AI can now write, reason, analyse, and create at a level that was science fiction five years ago.

The question everyone keeps asking is "when does AGI arrive?" But that's actually the wrong question. The more useful one is: when does AI get good enough, cheap enough, and reliable enough to do your specific job? And for a growing number of people, that day has already come.

This article pulls together the latest data, expert forecasts, and real-world layoff numbers to give you a clear-eyed view of what's happening, what's coming, and which jobs are most — and least — at risk.

The AGI Timeline: Expert Forecasts Are Converging

Something remarkable has happened to AGI predictions over the past three years. They've collapsed inward. In 2020, the median expert forecast for AGI put it roughly 50 years away. By February 2026, the Metaculus forecasting community — which aggregates thousands of predictions and has a strong track record on near-term events — puts a 25% probability on AGI arriving by 2029 and a 50% chance by 2033.¹

The people building the technology are even more aggressive. At the 2026 World Economic Forum in Davos, Anthropic CEO Dario Amodei said AGI would "likely occur within a few years," pointing to 2027 as plausible. DeepMind's Demis Hassabis was more cautious, maintaining roughly a 50% chance by the end of the decade.²

Prediction markets tell a slightly more sceptical story. Polymarket puts just a 9% probability on OpenAI achieving AGI by 2027, while Kalshi gives a 40% chance by 2030.² The money, it seems, is a bit more cautious than the rhetoric.

The Definition Problem

Much of the disagreement comes down to what AGI actually means. Current AI models already have an IQ above 130, speak 20+ languages, write code in most programming languages, and can explain complex topics with remarkable clarity.⁴ In narrow terms, we already have superhuman AI. But reliability, long-horizon planning, and genuine novel reasoning remain elusive. The practical impact on jobs, though, doesn't require full AGI. It just requires AI that's good enough at enough tasks to make a human worker redundant — and that threshold is far lower.

The Jobs Picture: What's Actually Happening Right Now

Forget the theoretical forecasts for a moment. Here's what the numbers say is already happening.

That headline number, reported by outplacement firm Challenger, Gray & Christmas, only captures companies that explicitly cited AI when announcing layoffs.⁵ The real figure is almost certainly higher. Many companies fold AI-driven restructuring into broader "efficiency" or "reorganisation" language.

And the pace is accelerating. In the first two months of 2026, technology firms alone have cut 32,000 jobs, typically the leading indicator for broader workforce changes.⁵ Major companies have been explicit about the connection:

Amazon eliminated 14,000 corporate roles in late 2025, followed by 16,000 more in January 2026, with AI-driven restructuring cited as a factor. Salesforce cut 4,000 support roles as AI took over half of customer queries. Workday cut 8.5% of its workforce (roughly 1,750 jobs) to reallocate resources toward AI investment. IBM's AskHR system now handles 11.5 million interactions annually with minimal human oversight.⁵⁶

The Sceptical View Matters Too

Here's where it gets nuanced. A Gartner survey of 321 customer service leaders found that only 20% had actually reduced staffing because of AI. The remaining 80% reported steady headcount, with AI helping existing teams handle more volume.⁷ Gartner's analysts have gone further, predicting that half of companies that cut workers for AI will rehire them by 2027 under new titles — and companies like Klarna have already reversed course, bringing back human customer service staff.

A Harvard Business Review study from January 2026, surveying over 1,000 executives, found something revealing: most AI-linked layoffs are happening in anticipation of AI's impact, not because of proven performance gains.⁸ Companies are cutting staff because they believe AI will replace those roles — not because it already has. That's a meaningful distinction.

And Goldman Sachs' chief economist, Jan Hatzius, was blunt in a February 2026 assessment: massive AI investment contributed "basically zero" to US economic growth in 2025. The direct impact on measured GDP was a negligible 0.2%.⁹

"We don't actually view AI investment as strongly growth positive. We think there's been a lot of misreporting of the impact that AI investment had on GDP growth in 2025, and it's much smaller than it's often perceived."— Jan Hatzius, Chief Economist, Goldman Sachs (February 2026)

MIT economist Daron Acemoglu takes an even more conservative position, estimating that only about 5% of tasks will be profitably automated by AI within the next decade, translating to a GDP boost of roughly 1.1% — "nontrivial but modest."¹⁰

Which Jobs Are Hit First — And Which Are Safe?

The displacement isn't happening evenly. It follows a clear pattern: AI targets tasks that are repetitive, rules-based, and digitally mediated. The more of your job that fits that description, the more exposed you are.

The Displacement Waves

Sources: McKinsey, Goldman Sachs, WEF Future of Jobs Report 2025, Careery AI Resistance Score framework.⁶¹¹

The "Junior Crisis"

The most concerning pattern isn't mass layoffs at the top. It's the disappearance of entry-level roles — the on-ramp jobs that train the next generation. Employment among college graduates aged 22–25 in AI-exposed fields has dropped 13% since late 2022.⁶ Entry-level job postings have fallen 15% year-on-year.

This matters because those junior roles have traditionally been how people learn an industry. If AI handles the work that used to go to graduates, where does the next generation of senior professionals come from? As Bank of England Governor Andrew Bailey warned in December 2025, AI is "likely to displace British workers" with particular concern about dampening the pipeline of young talent.¹²

In 2025, 40% of young university graduates chose careers in trades like plumbing, construction, and electrical work — roles that can't be automated.⁶ That's a rational response, but it represents a remarkable shift in how young people view the value of a knowledge-work degree.

The UK Picture: Harder Hit Than You'd Think

Britain is getting hit harder than comparable economies, and the numbers are stark.

Morgan Stanley's analysis of nearly 1,000 businesses across five AI-exposed industries found that while UK companies were eliminating roles at about the same rate as the global average, they were creating far fewer new AI-related positions to offset the losses.¹³ The result: Britain recorded the largest net employment decline among the countries studied.

The broader context isn't encouraging either. UK unemployment is at its highest level in four years. Job vacancies have fallen roughly a third since 2022 — around half a million fewer roles. And job adverts for occupations with high AI exposure have dropped 38% since 2022.¹²

The IPPR estimates that up to 7.9 million UK jobs could be exposed in a "second wave" of AI automation, with 1.5 million at risk in a worst-case near-term scenario. Women, younger workers, and lower-paid employees face disproportionate risk.¹⁴

London's mayor has flagged particular concern for the capital's white-collar economy. At his annual Mansion House speech, Sadiq Khan warned about AI's impact on finance, law, consulting, and the creative industries — sectors where London is disproportionately concentrated. A City Hall poll found 56% of London workers expected AI to affect their job in 2026.¹³

The Marketing Angle

For those of us in marketing specifically, the data is more nuanced than the headlines suggest. A survey of 1,000 UK marketing professionals found that 54% plan to expand internal marketing teams in 2026, while just 6.7% expect team size to decrease.¹⁵ AI's primary use in marketing is data analysis and insights (18.9%), followed by personalisation (13.2%) and campaign optimisation (11.5%). Content creation — despite the media attention — attracted only 12.1% adoption.

The message: marketing roles are transforming but not disappearing. The risk sits squarely with execution-level tasks (templated content, basic SEO articles, routine social posts), not strategy, relationships, or creative direction.

The Economics: Hype vs. Reality

The gulf between AI investment and AI returns is one of the underreported stories of 2026. The five largest US tech companies are collectively expected to spend as much as $700 billion on AI infrastructure this year.⁹ And yet:

Over 80% of companies report no productivity gains from AI so far despite billions in investment. Only about 1% of businesses believe their AI efforts have reached maturity. And 90% of the more powerful "vertical" AI use cases — where AI fully automates specific business processes — remain stuck in pilot mode.¹⁶

This doesn't mean AI won't deliver. Goldman Sachs has estimated that generative AI could raise global GDP by 7% (roughly $7 trillion) over a 10-year period.¹⁷ McKinsey puts the figure even higher, at $13 trillion in additional global output by 2030, or $2.6 to $4.4 trillion annually from generative AI alone.¹⁸

But the contrast between these projections and the current reality — where AI's contribution to US GDP growth was "basically zero" in 2025 — should give us pause. The economic benefits are likely coming. They just haven't arrived yet. And the labour market disruption is happening regardless.

What Does This Mean For You?

If you've read this far, you're probably wondering where you personally stand. Here's the honest assessment.

If you're in a high-risk category

The window for transition is 12–24 months for the most exposed roles. That's not a guess — it's the consensus view from McKinsey, WEF, and multiple workforce analyses. If your job is primarily data entry, Tier 1 customer service, basic content production, or routine administrative work, the time to reskill is now, not when the redundancy notice arrives.

If you're in a medium-risk category

Your role isn't going to disappear overnight, but it will change substantially. The people who thrive will be those who use AI as a multiplier — handling the routine 60% of their tasks with AI and focusing their human effort on the strategic, creative, and relationship-driven 40% that machines can't yet touch. Professionals with AI skills now command salaries up to 56% higher than peers in identical roles without those skills.¹²

If you're in a low-risk category

Skilled trades, healthcare, education, and roles requiring physical presence or deep human connection remain well-protected. The irony is that the jobs our society has traditionally undervalued — caring, building, fixing things with your hands — are the ones AI can't easily replicate.

The universal advice

Learn to work with AI, not against it. The data consistently shows that AI augments more roles than it eliminates, and the fastest-growing job categories are all AI-adjacent: AI engineer, machine learning researcher, AI consultant, AI governance specialist. The people who lose out aren't those whose jobs AI can do — it's those who refuse to learn how to leverage AI in their existing role.

The 2026–2030 Outlook

Pulling all of this together, here's the most likely trajectory:

2026–2027: The "junior crisis" intensifies. Entry-level compression accelerates across knowledge work. Agentic AI pilots hit 37% of large firms. Customer service, data entry, and basic content roles see significant reductions. Net job creation remains positive globally but distribution is highly uneven.¹¹

2028–2029: Widespread agent adoption. AI agents coordinate across enterprise systems, moving from task-level automation to managing entire workflows. Multi-agent systems become standard at large companies. The gap between AI-fluent and AI-illiterate workers becomes a chasm.

2030: A new equilibrium begins to form. The WEF's projected net gain of 78 million jobs should be materialising, but predominantly in roles that didn't exist in 2024. The workforce looks fundamentally different — fewer but more leveraged roles, higher productivity per worker, and an economy where the ability to direct and supervise AI systems is as basic as knowing how to use a spreadsheet.

The optimists and the pessimists are both partly right. AI will destroy millions of jobs. It will also create millions more. The problem is that the destruction and creation don't happen to the same people, in the same places, at the same time. That's the transition challenge — and it's the part that governments, businesses, and individuals need to take far more seriously than they currently are.

"AI job displacement statistics are not a prophecy of mass unemployment. They are a warning about misalignment during rapid transition."

The technology isn't the problem. The speed is.