Will AI replace your job? The fear of technology replacing jobs is not new. William Lee patented the mechanised loom for stocking production in the 16th century. In early 19th-century England, a group of textile workers, the Luddites, rioted and destroyed machinery that was automating the production of textiles, threatening their livelihoods. Often, technological advances create new jobs as well as rendering some obsolete, but the process can still be disruptive to those affected. There are not many people building slide rules these days, an industry that lasted for three centuries. In 1950 around 350,000 people were employed as switchboard operators working for phone companies, and another million in the same role at offices. The introduction of mechanised switchboards completely removed the need for these jobs, and by 1978, the last mechanical switchboards were replaced.
The latest technological advance to threaten jobs has been the advent of artificial intelligence (AI), especially the generative AI tools like ChatGPT that appeared in late 2022. Clearly, the impact of such tools will vary greatly depending on the profession. A chatbot is not going to cut your hair, paint your house, or build you a fence anytime soon. A July 2025 Microsoft study identified 40 careers that would have a high level of exposure to AI, such as interpreters, copywriters, ticket agents and customer service staff. Even photographic models are not safe in an age where a realistic stock photo of a person is just a prompt away.
On the other hand, people who operate pile drivers, dredgers, and logging equipment need not be overly concerned. Roofers, dishwashers and housekeepers can also rest easy.
McKinsey reckons that 80% of the world’s global companies were using AI in at least one business function by July 2025. This is not to say that all these efforts are actually succeeding. The same McKinsey report found that just a fifth of companies report a tangible effect on their earnings, and a comprehensive 2020 MIT study of hundreds of AI projects revealed that 95% fail to deliver any benefit whatever. Nonetheless, the McKinsey study observed that job vacancies fell by 43% between May 2022 and May 2025 in the USA. A look at on-line job adverts found the effect was much worse in AI-exposed jobs (a 38% drop) than in ones with low exposure to AI (a 21% drop). An in-depth August 2025 study by Stanford University confirmed this. The Stanford study found that AI seemed to be having an effect on the US job market, specifically in entry-level jobs, typically taken up by recent graduates. Since late 2022, employment for 22–25 year-olds in AI-exposed jobs (like software development & customer service) dropped 13-20%. Elsewhere, employment levels were either stable or grew. Young graduates are facing a tricky job market.
It is not clear whether this trend will continue, or whether many of the pauses in jobs have been in anticipation of savings from AI that have yet to really appear. Swedish Fintech company Klarna sacked 700 customer service workers and replaced them with AI, only to rehire them shortly afterwards when they saw the effect that the change was having on customer satisfaction.
One of the common use cases for generative AI is in software development. Although large language models (LLMs) like Claude Sonnet and tools like Microsoft Copilot cannot carry out all coding tasks, they can certainly deliver workable code in simpler use cases. This has led to a concern that junior software developers’ jobs are very much at risk. Of course, if you get rid of junior software developers, how will you develop the senior ones still needed to design systems and to fix the complex problems that LLMs cannot handle? Amazon’s AWS CEO Matt Garman stated that “replacing junior developers with AI was the dumbest thing I’ve ever heard.” for this reason.
Workers can react to the potential threat of AI in a number of ways. As noted, many jobs will be almost entirely unaffected. However, if you are in an area in the AI crosshairs, such as a graphic designer or translator, then you will need to think carefully about how you will continue to prosper. A world where an AI can do a decent translation of a document, or produce a stock photo or logo, within seconds, is clearly going to be a challenge, at the least, for people working in those areas. Most jobs will be in the middle of the spectrum, with some impact likely from AI. In such cases, workers should educate themselves about AI and how to work with it to make themselves more productive and effective. There are plenty of limitations of LLMs, such as their propensity to hallucinate and behave in ways that are inconsistent. Learning what AI is good at and what it is not good at will, in itself, be an important skill. It is likely that many jobs will be augmented by AI tools rather than replaced altogether. AI tools can be persuasive but lack empathy, as the increasing list of human tragedies caused by AI chatbot therapists mounts. Hence, workers can focus on learning and enhancing skills like this, which LLMs lack. We should also remember that we are barely three years into generative AI being available on a broad basis, and that it takes time for any new technology to be understood and adapted to. A PwC 2025 study found that AI can make workers more productive, with much faster revenue growth in industries most exposed to AI, and twice the growth in salaries in those industries. An example of how humans have learned to work with AI can be seen in radiology, where AI scans medical images and highlights potential issues in the scans, such as possible cancer, for the radiologist to review. The use of AI to augment rather than replace jobs is likely to happen in many fields.
It seems likely that AI will have a significant effect on some industries and careers, while many, such as those involving manual labour, will remain almost entirely unaffected. However, AI is likely to have at least some impact on a wide range of jobs, and so workers and employers need to figure out how to best utilise this technology so that workers can become more productive and effective. As the part-human, part-machine character Seven of Nine said in the TV series Star Trek: “We will adapt”.
