Fifteen “Hellish” Years Ahead: Will AI Trigger Unemployment and Crisis?

A Summer of Releases That’ll Make Your Head Spin

The first week of August alone was a fireworks display of AI announcements.

OpenAI rolled out the shiny new GPT-5 and a family of open-source language models, GPT-OSS (Open Source Scale).
On the same day, Elon Musk promised that xAI’s fifth-generation Grok would land by the end of 2025—hot on the heels of their Grok Imagine, an image and video generator powered by the Aurora model.
Anthropic upgraded Claude Opus to version 4.1, now a heavyweight rival to GPT-5, excelling in reasoning, complex text processing, and multimodal tasks.
Google DeepMind unveiled Genie 3, which users call a leap from content creation to generating living, interactive worlds from simple prompts.
ElevenLabs dropped IIElevenMusic—a service that churns out instrumental tracks and vocal compositions up to four minutes long.

The possibilities are exploding. Even a non-techie can now whip up results in minutes that used to take teams weeks or months. The flip side? The unemployment market is growing—and it’s already here. New grads are staring at a grim choice: no jobs or a mad dash to retrain. The thrill of AI’s capabilities is increasingly giving way to anxiety, doubt, and fear—am I even needed anymore?

Take me, a frontend developer. I’m already building full-blown apps—database, backend, the works—solo, with ChatGPT as my wingman. And I’m floored: no more days spent poring over docs, haunting forums, or pinging colleagues. It’s all sorted in minutes. Even this project you’re reading? I built and run it myself. AI is rewriting the rules—of work, of life.

Mo Gawdat’s Forecast: “Hell” Starts in 2027

Mo Gawdat, former Google exec and an early prophet of AI’s meteoric rise, now warns of a “hellish” 15-year stretch starting in 2027.

Now I believe that we are going to hit a short-term dystopia. There’s no escaping that.

I count it exactly as 12 to 15 years. I believe the beginning of the slope will happen in 2027. We will see signs in 2026. We’ve seen signs in 2024, but they will escalate next year and then a clear slip in 2027.

— Mo Gawdat

Creeping automation will erase most “white-collar” jobs. In a podcast with The Diary of a CEO (watch here), he’s blunt: AI won’t just replace call-center operators or assistants—it’s coming for developers, podcasters, even CEOs. His example? A startup called Emma.love, where tasks once handled by 350 people are now managed by three.

I want to believe he’s exaggerating. But the COVID lockdown feels like it was just a stress test for what’s coming.

Gawdat predicts the first painful signs will hit in early 2026, with the peak in 2027. He points to seven areas AI will reshape—his acronym FACE RIPS:

Freedom
Accountability
Connection
Equality
Reality
Innovation (as meaning)
Power (its distribution)
Society & economics

You’re going to have a massive concentration of power… you’re going to see a world with a lot of control, a lot of surveillance, a lot of forced compliance.

— Mo Gawdat

In the Danger Zone—and Beyond

The first jobs to go are those where value lies in repetition, not uniqueness. AI is already clearing out call-center operators, secretaries, content managers, and even some IT roles. It hits hardest where tasks involve mechanically processing requests, searching for info, or churning out standard reports or texts.

Let’s be clear: AI doesn’t “kill” entire professions. It carves out the routine—the dull, templated stuff people did on autopilot. But with that routine often goes half the workforce, because companies no longer need a full staff for the few creative or non-standard tasks.

Journalists and bloggers? That’s a separate story. Sure, AI can spit out a press release or rewrite a news story from ten sources in seconds. But it won’t wade through mud at a Brooklyn protest, stand in the rain outside a Manhattan courthouse, or catch the tremble in someone’s voice. Empathy, the ability to filter a story through your own lens, a unique perspective, boots-on-the-ground work—that’s still a human privilege. And it’s what people will pay for. Everything else? Machines already do it better and faster.

The jobs safest from AI, for now, involve live human contact—physical or emotional—or quick decisions in unpredictable settings:

Nurses, caregivers, physiotherapists
Hairdressers, manual therapists, chefs
Builders, rescuers, teachers, coaches

An algorithm can suggest, but it won’t pat a shoulder, turn over a heavy patient, or inspire confidence with a single glance.

Also safe—for now—are roles blending high uncertainty, creativity, and complex interpersonal skills:

Psychotherapists, social workers
Designers, field researchers
Strategists, crisis managers, team leaders, mentors

Where you need to inspire, persuade, read a room, or act on instinct, machines remain assistants, not leaders.

Voices of Experts and Researchers

MIT Media Lab (2025) warns that over-reliance on tools like ChatGPT can dull brain activity, creativity, memory, and critical thinking—especially in younger users (read more). This isn’t about fetching coffee or filing papers—it’s about the deep cognitive skills once central to analytical and text-heavy professions.

Meanwhile, MIT Sloan (2025) highlights that skills AI can’t replicate—empathy, ethics, creative imagination, the ability to inspire—are becoming more valuable (read more).

McKinsey projects that by 2030, automation will disrupt up to 30% of the workforce—one in three people will face turbulence (read more). Some will retrain, some will flounder, some will be left behind.

Harvard adds: those at greatest risk are workers whose tasks are 60–80% repetitive. Anything that can be scripted, AI will do faster (read more).

Sectors already feeling the heat:

Admin and office management: from calendars to email filters.
Finance and accounting: numbers and templates are already automated.
Media: news blurbs, press releases, SEO copywriting—where speed, not soul, matters.
IT: especially front- and backend roles stuck in repetitive loops. If your code can be templated, it’ll be generated.

IT on the Edge: Is Frontend Dead? Or Backend?

A frontend dev scrolls through their feed and sees a fresh grad lamenting: “If you’re only in frontend, just retrain—your job’s toast!” But cooler heads, like Josh W. Comeau, push back. He argues frontend isn’t dying—it’s evolving. His take: AI isn’t a replacement, it’s an amplifier. Like a drill in a craftsman’s hands, not a machine gun in a rookie’s.

Reality check: GPT-5 and its peers are nailing boilerplate code, simple landing pages, docs, bug fixes, and tests. But a serious app—architecture, performance, security, UX? That’s another league. Frontend is complex: it’s not just HTML and CSS, but thousands of files, state management, accessibility, style, performance, and product logic.

A Dev.to poll by Sacha Greif shows 91% of developers use AI for coding help, but 61% rewrite AI-generated code—poor readability, holes, bugs. Bottom line: about 28% of code comes from AI, but only a fraction makes it to production without human tweaks. AI’s output is a draft, not a final product.

X posts scream: “Once AI learns to write clean, efficient code and self-check, it’s game over for devs.” But for now, AI lacks full context, doesn’t hold product strategy in its head, and isn’t accountable. It’s a tool, not a replacement.

Other Industries? Not Safe, But Calmer

Admin roles—secretaries, office managers, coordinators—are fading fast. Emails, meetings, internal requests, approvals? AI bots handle those. Humans are left for conflicts and situations needing common sense, not rules.

Finance and accounting? Algorithms are right at home with account reconciliation, reporting, taxes. The danger: companies are slashing entire departments, keeping just a few overseers.

Media? News snippets, press releases, SEO texts, marketing emails—AI nails speed and accuracy. But investigative reporting, field work, or columns with a personal voice? Still human territory. The winners combine automation with a human lens.

Journalists and Bloggers: How to Fight Off AI

An algorithm can churn out a dry news piece or a tidy press release in seconds. But it won’t brave the rain for a stranger’s comment, smell smoke at a fire, or grasp the scale of grief.

In an era of cookie-cutter AI content flooding feeds, value lies in what only humans can deliver:

Personal experience and an insider’s perspective.
Empathy and the ability to truly listen.
Fieldwork—where there’s no API or dataset.
A voice that either hooks or repels from the first line.

Paradoxically, as automation surges, hands-on journalism and personal blogs could see a renaissance—if writers stop racing machines for speed and volume.

Future Scenarios

AI’s at the doorstep. The question is which door it’ll walk through.

Optimistic: Automation lifts the grunt work, freeing people for what they love—creativity, science, art. Professions don’t vanish; they evolve. Workweeks shrink, “work for life” overtakes “live to work.”

Pessimistic: Elysium in real life—an elite few control AI, reap its rewards, and gatekeep access to tech.

The elites are living on the moon. They don’t need peasants to do the work anymore… and they’ll decide who gets access to the machines that run the world.

— Mo Gawdat

He warns of a world where “the elites are living on the moon. They don’t need peasants to do the work anymore…”

Realistic: A hybrid. AI takes 80% of repetitive tasks, shifting value to human skills—critical thinking, creativity, trust-building, negotiation.

AGI in my mind… we will get to AGI in 2026, latest. AI is already smarter than me in everything. The real turning point will be self-evolving AI—that’s when the 250 IQ accelerates fast, and we get into an intelligence explosion. No doubt about it.

The only barrier between a utopia for humanity and AI and the dystopia we’re going through is a capitalist mindset.

— Mo Gawdat

The takeaway? Those who thrive will team up with AI, adapt, and use it as a partner, not an enemy.

What to Do Now

The question isn’t will AI replace me? It’s when and which parts of my work. Panic’s a waste. Action’s the plan.

Learn on the Fly
Learning isn’t linear anymore—school, college, job. It’s a loop: master one skill, add another, tweak a third. Courses, bootcamps, mentorship, real-world experiments. The ability to learn fast trumps any degree.

I used to do an hour a day of reading… Now I do four hours a day. It is impossible to keep up—the world is moving so fast.

— Mo Gawdat
Hone What Machines Can’t Touch
Empathy, negotiation, curiosity, ethics, truth-seeking. Anything requiring human connection, not algorithms. In a world of AI-generated content and decisions, value lies in staying human—bridging people, loving authentically, staying curious and honest.

The biggest skill that humanity will benefit from in the next 10 years is human connection… Not the smartest people, but the people that connect most to people are going to have jobs going forward.

I encourage people to question everything.

Magnify ethics so that the AI learns what it’s like to be human.

— Mo Gawdat
Make AI Your Tool
Integrate it into your workflow before it comes for your job. Automate the boring stuff, check code, speed up research, test ideas. Let AI handle what’s repetitive so you can focus on choice, taste, and courage. This isn’t cheating—it’s the new normal.

There will be two stages… the era of augmented intelligence… and then the era of machine mastery.

— Mo Gawdat

We’re in the human + AI phase now, but AI-first is coming. Be the one steering the partnership, not the one replaced.
Think Beyond Your Job
Your role is part of a massive system that’s shifting. Staring at your feet means missing the ground moving beneath you.

We need to rethink the entire model of work, economics, and how we distribute resources in a world where AI does most of the labor.

Extreme poverty could end for 10–12% of [military] spending… You can end world hunger for less than 4%.

— Mo Gawdat

The future hinges on a fork:

I call that the MAD spectrum—Mutually Assured Prosperity versus Mutually Assured Destruction.

— Mo Gawdat

Either we rebuild for shared good, or we get a tech-driven “every man for himself.”
Tactical Pause for IT
People scream “frontend’s dead!”—but “backend’s dead” gets equal hype. Reality’s simpler: AI eats routine tasks, junior-level repetition, docs, tests, basic landing pages. Architecture, thoughtful UX, security, scalability? Still human territory.

The replacement of physical manual labor is going to take four to five years before it’s possible at the quality of the AI replacing mental labor now… blue collar will stay longer.

— Mo Gawdat

Translation: you’ve got a few years to sharpen skills that resist automation and learn to orchestrate AI, not fight it.

The Bottom Line

AI won’t kill professions—it’ll reshape them. Clinging to the old “study, work, retire” model will leave you on the sidelines.

This isn’t just a tech shift; it’s a cultural overhaul—with new logic for value, power, and the meaning of work.
We can build mutually assured prosperity, where tech serves the common good, or slide into an “every man for himself” world where AI becomes a tool for control and enrichment of a tiny elite.

The question isn’t what AI will do. It’s what we will do—and whose future we’ll build: one for everyone, or just for those who can afford the “token” to get in.