AI at Work: How many jobs are really at risk?
Artificial intelligence is reshaping the labour market, but the landscape is not a simple one-way street. A Microsoft study examining 200,000 conversations with its Copilot chatbot argues AI can handle many tasks across dozens of roles, yet company leaders and workers warn that the picture is nuanced. The result is a growing debate: which jobs are most exposed to automation, and which are shielded—at least for now.
Real stories behind the numbers
Joe Turner, a 38-year-old writer, embodies the contrast. He lost roughly 70% of his client base to AI-driven processes in two years, wiping out around £120,000 of annual revenue. For Turner and others in creative fields, AI can draft, edit, and even pitch ideas at astonishing speed. Yet the human touch—the nuance, personality, trust, and nuanced judgment—remains hard to replicate.
Microsoft’s analysis suggests AI can complete at least 90% of work for historians and coders, 80% for salespeople and journalists, and 75% for DJs and data scientists. But the same study also shows many roles with partial automation, where AI handles routine components while humans steer direction, strategy, and complex decision-making.
40 high-risk roles versus 40 that resist automation
In the Money team’s conversations with experts, researchers, and workers, a list of vulnerable occupations emerged. Customer service, financial advisory, and product promotion sit high on the exposure spectrum, with job tasks that are repetitive or rule-driven often more susceptible to AI augmentation. Equally, the study highlights occupations where AI can perform only a portion of duties, suggesting a future where humans and machines collaborate rather than a wholesale replacement.
Conversely, roles tied to relationships, nuanced judgment, or bespoke problem-solving show greater resilience. Specialists who rely heavily on empathy, high-stakes decisions, or complex human interactions—areas where AI still struggles—tend to be more shielded. The argument isn’t destiny, however: availability of AI tools, business models, and macroeconomic factors can shift the balance quickly.
Is the future all doom and gloom or a call to reskill?
Industry voices are split. Some executives tout automation as a productivity boon that creates new work in fields like AI maintenance, data governance, and design of human-centric AI experiences. Others worry that headline forecasts of mass displacement ignore the reality that many firms are pausing hiring or delaying layoffs while they assess how best to deploy AI.
A senior AI consultant cautions against reading the data as a snap mandate for immediate job losses. He notes that hiring freezes and cautious experimentation with AI can mask an underlying drift—where job profiles gradually shift, not vanish overnight. The UK’s job postings show growth in AI-related roles, but at a slower pace than some sectors. The takeaway is less about inevitability and more about adaptation.
What workers can do to navigate the AI shift
Experts advise proactive skill-building. The World Economic Forum cites AI competency as an increasingly core skill, and LinkedIn data shows AI skills proliferating in professional profiles. For workers: learn to collaborate with AI, identify tasks that can be automated, and reframe your role to emphasize areas where human judgment, creativity, and relationship-building matter most.
For freelancers like Turner, diversification and client communication are critical. Some clients are returning to demand human-authored content, while others are exploring hybrid approaches that blend AI-generated drafts with professional editing and storytelling. The headline warning—AI could displace tasks—often coexists with an opportunity: AI can unlock new, higher-value work for those who evolve with the technology.
Bottom line
AI is not a universal job-killer, but it is a powerful force for task-level automation. The most important safeguard for workers is ongoing upskilling and a willingness to adapt. As firms experiment with AI, the human ability to think creatively, relate to others, and exercise judgment remains the differentiator that technology struggles to mimic.
