As Moravec predicted, AI is taking your office jobs but will let you live if you do the dishes and laundry… for now

The conversation about AI replacing human jobs is no longer entirely hypothetical. We are well past the era of speculative think pieces about whether machines might someday take over our work.

That future is already here – and it’s showing up in surprising places.

What makes this moment different from past automation waves is that AI is now targeting roles once considered safe: creative, analytical, and highly paid white-collar jobs.

The “robots” in question are not assembly line arms, but digital minds like ChatGPT, Claude, and GitHub Copilot – systems capable of generating text, code, images, and even making strategic decisions.

And while some industry leaders claim that AI merely “assists” workers, the evidence from company payrolls tells a different story.

The new office ‘colleagues’

If you work in an office today, you probably already have an AI assistant – whether you asked for one or not.

• ChatGPT (OpenAI) can draft documents, answer questions, write code, and provide on-demand research.
• Claude (Anthropic) offers long-form analysis, summarisation, and structured reasoning with a more “chatty” tone.
• GitHub Copilot uses AI to autocomplete code, detect errors, and write functional scripts in seconds.
• Artisan AI agents and similar tools act as autonomous “digital employees,” running customer service workflows, content moderation, or data entry with minimal oversight.

These systems are deployed across marketing departments, HR teams, software development groups, and even legal practices. In the short term, they speed up work and reduce costs. In the long term, they reduce the need for junior staff – once the natural training ground for future managers.

Companies already replacing humans with AI

Over the past 18 months, AI adoption has coincided with layoffs in sectors that once seemed untouchable.

• TikTok recently laid off 150 trust and safety moderators in Berlin – about 40 percent of the team – replacing their work with AI moderation systems and cheaper outsourced labour.
• Tata Consultancy Services (TCS), one of the world’s largest outsourcing companies, cut more than 12,000 jobs in India as tasks like coding, testing, and customer support shifted to AI systems.
• US tech giants – Amazon, Microsoft, Walmart – have cut roughly 130,000 tech roles this year, with AI automation cited as a key driver of “efficiency”.
• Klarna, Duolingo, Cisco, UPS, and Intuit have each replaced specific human functions – from marketing copy to logistics scheduling – with AI.

According to Challenger, Gray & Christmas, AI has been directly linked to more than 27,000 US job losses since 2023, as reported by CBS. Meanwhile, Business Insider reports that over 10,000 job cuts this year alone are explicitly attributed to AI.

The pain is not evenly spread – Goldman Sachs found unemployment among young tech workers (ages 20 to 30) has jumped three percentage points in just months.

Some leaders are blunt about what’s coming. Anthropic’s CEO predicts that up to half of entry-level white-collar jobs could vanish within five years, potentially pushing US unemployment to 20 percent.

JPMorgan analysts warn of a “jobless recovery”, where economic growth returns but hiring for office roles does not.

The paradox of ‘skills at risk’

It’s tempting to think AI is targeting only low-level work. But the technology has proved surprisingly adept at complex, “skilled” tasks like writing, software engineering, and medical diagnostics – while struggling with chores like folding laundry or washing dishes.

This reversal of expectations is known as Moravec’s Paradox: the things humans find intellectually hard are often easy for machines, while the physical, sensory tasks we find effortless are fiendishly difficult for AI.

That’s why stock trading has been largely automated for years, while your dishwasher still can’t load itself. It’s also why someone who dreamed of AI doing her cleaning so she could focus on creative writing now finds herself with the opposite: AI writes her articles while she does the washing up.

Are software engineers obsolete?

Not yet – but their role is shifting. Tools like GitHub Copilot can produce working code from plain-language prompts, making them valuable for rapid prototyping and bug fixing. Entry-level coding work – the first rung on the career ladder – is the most vulnerable.

Experienced engineers are still needed for architecture, integration, and oversight. But the ladder itself is shrinking. New graduates may find fewer opportunities to learn on the job before AI takes over the repetitive tasks they once cut their teeth on.

The ‘prompt engineer’ hype

In the early days of large language models, companies offered six-figure salaries for “prompt engineers” who could coax optimal outputs from AI.

The reality? Yes, some prompts are simple. Others require technical knowledge, understanding of model behaviour, and skill in combining multiple tools.

Still, as AI interfaces improve, prompt writing will likely become a standard part of many jobs, not a stand-alone career.

How far could AI go?

Studies vary, but the broad consensus is that most jobs will be changed by AI – some more radically than others.

OpenAI estimates that 80 percent of US workers could have at least 10 percent of their tasks affected by large language models, with 19 percent facing disruption to more than half their daily work.

Whether this leads to mass unemployment or a reimagined job market depends on political and corporate choices. Automation could make societies richer, healthier, and more innovative – or it could concentrate wealth and power while sidelining millions of workers.

A future worth planning for

The lesson from the past year is clear: AI in the workplace is not a distant threat – it’s an active force reshaping payrolls today. If history is a guide, new kinds of work will emerge, but the transition will be messy.

Governments, businesses, and individuals will need to rethink education, career planning, and social safety nets.

And maybe – just maybe – we’ll figure out how to make a robot that can do the laundry before it learns to replace us all.

Visit Us: Robotics and Automation