Ironically, coding and software engineering jobs are among the most vulnerable to the spreading of AI. While there are expectations that AI will increase productivity and streamline routine tasks with many programmers and non-programmers likely to benefit, some coders confess that they are becoming overly reliant on AI suggestions (which undermines problem-solving skills).

Anthropic, one of the leading developers of generative AI systems, recently launched an Economic Index based on millions of anonymised uses of its Claude chatbot. It reveals massive adoption of AI in software engineering: “37.2% of queries sent to Claude were in this category, covering tasks like software modification, code debugging, and network troubleshooting”.

AI is also outperforming humans in a growing array of medical imaging and diagnosis roles. While doctors may not be replaced outright, support roles are particularly vulnerable and medical professionals are getting anxious. Analysts insist that high-skilled jobs are not at risk even as AI-driven diagnostic tools and patient management systems are steadily being deployed in hospitals and clinics worldwide.

Meanwhile, the creative sectors also face significant disruption as AI-generated writing and synthetic media improve. The demand for human journalists, copywriters, and designers is already falling just as AI-generated content (including so-called “slop”: the growing amount of low-quality text, audio and video flooding social media) expands. And in education, AI tutoring systems, adaptive learning platforms, and automated grading could reduce the need for human teachers, not only in remote learning environments.

Arguably the most dramatic impact of AI in the coming years will be in the manufacturing sector. Recent videos from China offer a glimpse into a future of factories that run 24/7 and are nearly entirely automated (except a handful in supervising roles). Most tasks are performed by AI-powered robots and technologies designed to handle production and, increasingly, support functions.

Unlike humans, robots do not need light to operate in these “dark factories”. CapGemini describes them as places “where raw materials enter, and finished products leave, with little or no human intervention”. Re-read that sentence. The implications are profound and dizzying: efficiency gains (capital) that come at the cost of human livelihoods (labor) and rapid downward spiral for the latter if no safeguards are put in place.

Some have confidently argued that, as with past technological shifts, AI-driven job losses will be offset by new opportunities. AI enthusiasts add that it will mostly handle repetitive or boring tasks, freeing humans for more creative work — like giving doctors more time with patients, teachers more time to engage with students, lawyers more time to concentrate on client relationships, or architects more time to focus on innovative design. But this historical comfort overlooks AI’s radical novelty: for the first time, we’re confronted with a technology that is not just a tool but an autonomous agent, capable of making decisions and directly shaping reality. The question is not just what we can do with AI, but what AI might do to us.

AI will certainly save time. Machine learning already interpretsscans faster and cheaper than doctors. But the idea that this will give professionals more time for creative or human-centered work is less convincing. Already doctors are not short on technology; they are short on time because healthcare systems prioritise efficiency and cost-cutting over “time with patients”. The rise of technology in healthcare has coincided with doctors spending less time with patients, not more, as hospitals and insurers push for higher throughput and lower costs. AI may make diagnosis quicker, but there is little reason to think it will loosen the grip of a system designed to maximise output rather than human connection.

Nor is there much reason to expect AI to liberate office workers for more creative tasks. Technology tends to reinforce the values of the system into which it is introduced. If those values are cost reduction and higher productivity, AI will be deployed to automate tasks and consolidate work, not to create breathing room. Workflows will be redesigned for speed and efficiency, not for creativity or reflection. Unless there is a deliberate shift in priorities — a move to value human input over raw output — AI is more likely to tighten the screws than to loosen them. That shift seems unlikely anytime soon.

AI’s uneven impacts

AI’s impact on employment will not be felt equally around the world. It will impact different countries differently. Disparities in political systems, economic development levels, labour market structures and access to AI infrastructure (including energy) are shaping how regions are preparing for and are likely to experience AI-driven disruption. Smaller, wealthier countries are potentially in a better position to manage the scale and speed of job displacement. Some lower-income societies may be cushioned by the disruption owing to limited market penetration of AI services altogether. Meanwhile, high and medium income countries may experience social turbulence and potentially unrest as a result of rapid and unpredictable automation.

The United States, the current leader in AI development, faces significant exposure to AI-driven disruption, particularly in services. A 2023 study found that highly educated workers in professional and technical roles are most vulnerable to displacement. Knowledge-based industries such as finance, legal services, and customer support are already shedding entry-level jobs as AI automates routine tasks.

Technology companies have begun shrinking their workforces, using that also as signals to both government and business. Over 95,000 workers at tech companies lost their jobs in 2024. Despite its AI edge, America’s service-heavy economy leaves it highly exposed to automation’s downsides.

Asia stands at the forefront of AI-driven automation in manufacturing and services. It is not just China, but countries like South Korea that are deploying AI in so-called “smart factories” and logistics with fully automated production facilities becoming increasingly common. India and the Philippines, major hubs for outsourced IT and customer service, face pressure as AI threatens to replace human labour in these sectors. Japan, with its shrinking workforce, sees AI more as a solution than a threat. But the broader region’s exposure to automation reflects its deep reliance on manufacturing and outsourcing, making it highly vulnerable to AI-driven job displacement in a geopolitically turbulent world.

Europe is taking early regulatory steps to manage AI’s labour market impact. The EU’s AI Act aims to regulate high-risk AI applications, including those affecting employment. Yet in Eastern Europe, where manufacturing and low-cost labour underpin economic competitiveness, automation is already cutting into job security. Poland and Hungary, for example, are seeing a rise in automated production lines. Western Europe’s knowledge-based economies face risks similar to those in America, particularly in finance and professional services.

Oil-rich Gulf states are investing heavily in AI as part of diversification efforts away from a dependence on hydrocarbons. Saudi Arabia, the UAE, and Qatar are building AI hubs and integrating AI into government services and logistics. The UAE even has a Minister of State for AI. But with high youth unemployment and a reliance on foreign labour, these countries face risks if AI reduces demand for low-skill jobs, potentially worsening inequality.

In Latin America, automation threatens to disrupt manufacturing and agriculture, but also sectors like mining, logistics, and customer service. As many as 2-5% of all jobs in the region are at risk, according to the International Labor Organization and World Bank. And it is not just young people in the formal service sectors, but also human labour in mining operations, logistics and warehouse workers. Call centers in Mexico and Colombia face pressure as AI-powered customer service bots reduce demand for human agents. And AI-driven crop monitoring, automated irrigation, and robotic harvesting threaten to replace farm labourers, particularly in Brazil and Argentina. Yet the region’s large informal labour market may cushion some of the shock.

While most Africans are optimistic about the transformative potential of AI, adoption remains low due to limited infrastructure and investment. However, the continent’s rapidly growing digital economy could see AI play a transformative role in financial services, logistics, and agriculture. A recent assessment suggests AI could boost productivity and access to services, but without careful management, it risks widening inequality. As in Latin America, low wages and high levels of informal employment reduce the financial incentive to automate. Ironically, weaker economic incentives for automation may shield these economies from the worst of AI’s labour disruption.

No one is prepared

The scale and speed of recent AI developments have taken many governments and businesses by surprise. To be sure, some are proactively taking steps to prepare workforces for the transformation. Hundreds of AI laws, regulations, guidelines, and standards have emerged in recent years, though few of them are legally binding. One exception is the EU’s AI Act, which seeks to establish a comprehensive legal framework for AI deployment, addressing risks such as job displacement and ethical concerns. China and South Korea have also developed national AI strategies with an emphasis on industrial policy and technological self-sufficiency, aiming to lead in AI and automation while boosting their manufacturing sectors.

Notwithstanding recent attempts to increase oversight over AI, the US has adopted an increasingly laissez-faire approach, prioritising innovation by reducing regulatory barriers. This “minimal regulation” stance, however, raises concerns about the potential societal costs of rapid AI adoption, including widespread job displacement, the deepening of inequality and undermining of democracy.

Other countries, particularly in the Global South, have largely remained on the sidelines of AI regulation, lacking the awareness, capabilities or infrastructure to tackle these issues comprehensively. As such, the global regulatory landscape remains fragmented, with significant disparities in how countries are preparing for the workforce impacts of automation.

Businesses are under pressure to adopt AI as fast and deeply as possible, for fear of losing competitiveness. That’s, at least, the hyperbolic narrative that AI companies have succeeded in putting forward. And it’s working: a recent poll of 1,000 executives found that 58% of businesses are adopting AI due to competitive pressure and 70% say that advances in technology are occurring faster than their workforce can incorporate them.

Another new survey suggests that over 40% of global employers planned to reduce their workforce as AI reshapes the labour market. Lost in the rush to adopt AI is a serious reflection on workforce transition. Financial institutions, consulting firms, universities and nonprofit groups have sounded alarms about the economic impact of AI but have provided few solutions other than workforce up-skilling and Universal Basic Income (UBI). Governments and businesses are wrestling with a basic challenge: how to manage the benefits of AI while protecting workers from displacement.

AI-driven automation is no longer a future prospect; it is already reshaping labour markets. As automation reduces human workforces, it will also diminish the power of unions and collective bargaining furthering entering capital over labour. Whether AI fosters widespread prosperity or deepens inequality and social unrest depends not just on the imperatives of tech company CEOs and shareholders, but on the proactive decisions made by policymakers, business leaders, union representatives, and workers in the coming years.

The key question is not if AI will disrupt labour markets — this is inevitable — but how societies will manage the upheaval and what kinds of “new bargains” will be made to address its negative externalities. It is worth recalling that while the last three industrial revolutions created more jobs than they destroyed, the transitions were long and painful. This time, the pace of change will be faster and more profound, demanding swift and enlightened action.

At a minimum, governments must prepare their societies to develop a new social contract, prioritise retraining programs, bolster social safety nets, and explore UBI to help workers displaced by automation. They should also proactively foster new industries to absorb the displaced workforce. Businesses, in turn, will need to rethink workforce strategies and adopt human-centric AI deployment models that prioritise collaboration between humans and machines, rather than substitution of the former by the latter.

The promise of AI is immense, from boosting productivity to creating new economic opportunities and indeed helping solving big collective problems. Yet, without a focused and coordinated effort, the technology is unlikely to develop in ways that benefit society at large.