Is Your Job Safe From AI? The Ten Riskiest Occupations Right Now
The question on a lot of workers' minds isn't abstract anymore. It's simple: how likely is it that a machine takes over what I do for a living, and if it does, how soon?
The scale of the shift is hard to overstate. The World Economic Forum's 2025 Future of Jobs Report projects that 92 million jobs worldwide — roughly 8% of total global employment — will be displaced by 2030 as automation and AI adoption accelerate.
Closer to home, the picture is more modest but still significant. Signal 49, the research group formerly known as the Conference Board of Canada, estimated in January that AI-driven automation could wipe out up to half a million Canadian jobs by 2030 that would otherwise have existed.
Corporate Canada is already leaning in. A KPMG survey of the country's largest companies, published in November 2025, found that 93% have begun investing in or implementing AI in some form. That enthusiasm hasn't fully translated into results, though: only 31% say AI is truly embedded in their operations, and a striking 98% say they haven't yet seen a measurable return on their AI spending.
Not every job is equally exposed. According to the McKinsey Global Institute, the roles most likely to be handed over to machines share a common thread — they're predictable, repetitive, and built around collecting or processing data, tasks that software can now often do faster and more cheaply than a person.
To put a finer point on which occupations are most exposed, Anthropic, the AI company behind the Claude chatbot, has published research estimating what share of the tasks in dozens of occupations its own models can already perform. The following ten roles topped that list.
1. Computer programmers
Perhaps the least surprising entry on the list: Anthropic's data suggests roughly 75% of a computer programmer's day-to-day tasks can currently be handled by AI. Coding tools have improved so quickly that work which once took a human developer days or weeks can, in many cases, now be produced by an AI model in a fraction of the time — a gap that helps explain why this occupation sits at the top of the risk list.
2. Customer service representatives
Close behind, at around 70%, are customer service jobs. The tasks most exposed include answering customer questions, processing orders, and handling complaints — work that chatbots and AI-driven support systems increasingly manage without a human on the line. Researchers studying this shift note there's been no measurable rise in unemployment among workers in these roles so far, though there is early evidence that hiring has slowed slightly for younger workers, particularly those aged 22 to 25.
3. Data entry keyers and medical record specialists
About 67% of tasks in these roles can already be automated. For data entry keyers, that means reading documents and inputting information into computer systems — straightforward, rules-based work that AI handles comfortably. Medical record specialists face similar exposure, since AI systems can now collect, interpret, and code patient records with minimal human involvement.
4. Market research analysts and marketing specialists
Roughly 65% of this work is within reach of current AI systems, according to Anthropic's data — particularly the parts of the job that involve compiling findings into reports and explaining marketing data in plain language.
5. Sales representatives
For sales reps selling standard products to businesses — as opposed to technical or scientific goods, which require more specialized expertise — AI can reportedly handle about 62% of the job, including tasks like demonstrating products and processing orders.
6. Financial and investment analysts
Anthropic's Claude model is estimated to be capable of performing about 57% of this role, which centres on studying financial information and helping determine where money should be invested.
7. Software quality assurance analysts and testers
AI can reportedly handle around 52% of this job, including modifying software to fix bugs and improve performance — ironic, given that AI-written code is also part of what's driving demand for testing in the first place.
8. Information security analysts
An unglamorous but critical job, information security analysis is estimated to be about 49% automatable. AI systems are increasingly capable of scanning for vulnerabilities and helping ensure data stays secure as it moves through a network.
9. Computer user support specialists
These are the people who field questions about software glitches and hardware problems. Anthropic's research suggests AI can now handle about 47% of that troubleshooting work.
10. Tour guides
An honourable mention rather than a hard data point — Anthropic's study didn't formally assess this occupation — tour guides nonetheless face a new kind of competition. AI-enabled smart glasses can now deliver narrated tours directly into a wearer's ears. The Daily Mail has reported that tourists in New York are already testing an "AI live" feature on such glasses, letting users chat in real time with a built-in voice assistant while sightseeing. The hardware is loaded up, too, with a 12-megapixel camera, five microphones and two speakers built into the frame. At roughly $409 a pair, though, plenty of travellers may still prefer a human guide, at least for now.
Taken together, these ten roles share the same vulnerability: repetitive, structured tasks that are easy for a machine to learn and repeat. And the impact isn't just theoretical. Research out of Harvard Business School found that job postings in AI-exposed occupations have fallen 13% since ChatGPT's public launch in November 2022.
That's the sobering half of the story. The more encouraging half is that AI appears to be creating work even as it eliminates it. The same Harvard Business School analysis found that postings for jobs requiring analytical, technical, or creative skills grew by 20% over the same period.
The research, based on a review of American job postings from 2019 through early 2025, points to a broader pattern in how companies are adopting generative AI: rather than simply replacing workers, many employers appear to be restructuring roles around human-AI collaboration. Researchers even found new, AI-related skills appearing in job listings that hadn't existed in job descriptions before — skills that didn't fit neatly into old categories because the technology enabling them is so new.
It's worth noting that all of this data captures only the short-term picture. Whether these trends hold, accelerate, or reverse over the next decade remains an open question. What the Harvard researchers do say with more confidence is that workers looking to insulate themselves should focus on building "non-automatable skills, such as judgment and interpersonal communication skills."
Not everyone agrees the disruption is even underway yet. Yale research published in June, based on an analysis of U.S. labour data, concluded that the numbers "do not provide clear evidence of labor market disruption associated with AI" — a reminder that, for all the anxiety around the technology, the hard data on job losses remains contested.
For now, the reasonable takeaway isn't panic. It's adaptation. Whether or not a full-blown AI jobs crisis is on the horizon, the roles most exposed to automation today are also the ones where workers have the clearest signal to start building the skills machines still can't replicate.
Members Discussion