Employers are redesigning roles around AI, while public policy shifts toward training, infrastructure, and labour protections
THE UNIVERSAL RECORD
Sourced reporting. No opinions.
By Brad Socha | March 6, 2026 | 8:34 AM EST
Artificial intelligence is beginning to replace some job tasks, but the most consistent evidence still points to job transformation happening faster than outright job elimination. The International Labour Organization said in May 2025 that one in four jobs worldwide is potentially exposed to generative AI, while stressing that transformation, not replacement, is the more likely near-term outcome. The same ILO update said exposure is higher in high-income countries and highest in clerical work, while full automation remains limited by technical, organizational, and skills constraints.
That broader pattern is also reflected in employer surveys. The World Economic Forum’s Future of Jobs Report 2025, based on more than 1,000 large employers representing over 14 million workers, found that employers expect global macrotrends to create 170 million jobs and displace 92 million by 2030, for a net gain of 78 million. The same report says 39% of core job skills are expected to change by 2030, with AI and big data among the fastest-rising skills.
Where AI is replacing work first
The clearest replacement is happening at the task level, especially in repetitive digital work. The ILO says clerical occupations remain the most exposed because generative AI can handle many text-based, administrative, and documentation tasks. The organization also notes increasing exposure in highly digitized cognitive jobs in media, software, and finance.
The OECD reaches a similar conclusion. Its AI-and-work research says firms are still in an early adoption phase, but rapid technological progress, falling costs, and growing AI skills mean many advanced economies may be close to a broader AI expansion. The OECD also warns that automation is only part of the story: risks include loss of worker agency, privacy concerns, bias, and lack of transparency.
In workplaces already using AI heavily, the OECD reports that algorithmic management is spreading fast. It defines algorithmic management as using software, including AI, to fully or partly automate tasks traditionally performed by managers. In its surveys, 90% of U.S. managers said their firms use at least one algorithmic-management tool, and 76% of U.S. managers said their firms use ten or more of the fifteen tools covered in the survey.
What companies are trying to do long term
Large employers are not only testing AI assistants; many are moving toward operating models where humans work alongside software agents. Microsoft’s 2025 Work Trend Index says it expects every organization to be on a path toward becoming what it calls a “Frontier Firm” within the next two to five years. Microsoft describes this model as one built around hybrid teams of humans and agents, and says 81% of leaders expect agents to be moderately or extensively integrated into their company’s AI strategy within 12 to 18 months.
Alphabet’s public investor materials show a similar long-term focus on AI infrastructure. In its 2025 Q1 earnings call, Alphabet said its long-term investments in network and compute infrastructure are positioning it for AI demand, and highlighted continued investment in next-generation TPU and GPU capabilities. That suggests a long-range strategy centered on building the computing backbone needed for AI products, enterprise tools, and model deployment at scale.
Salesforce is already treating AI work as a measurable business output. In its fiscal 2026 results, the company defined an “Agentic Work Unit” as a discrete task executed by AI agents in production, such as resolving service cases, updating records, or triggering workflows. That is an important signal: some companies are moving from using AI as a productivity aid to counting AI systems as part of operational capacity.
Taken together, those company signals point to a long-term corporate goal that is less about eliminating all workers and more about reducing labour hours for routine work, increasing output per employee, and reorganizing teams around human oversight plus AI execution. That can still mean fewer roles in some functions, especially administrative support, basic content handling, and standardized service workflows, even if total firm output grows. This is an inference based on the company materials and labour-market evidence above.
What governments are doing about it
Governments are responding in two broad ways: building AI capacity and trying to prepare workers for labour-market disruption. The balance between those goals varies by country.
The European Commission is focused on scaling AI adoption and infrastructure. Its 2025 AI Continent Action Plan says the EU aims to mobilize €200 billion for AI development, finance up to five AI gigafactories, support 19 AI factories, and at least triple EU data-centre capacity over the next five to seven years. It also says only 13.5% of EU companies currently use AI, so part of the strategy is explicitly about increasing adoption in industry while also strengthening AI skills and talent.
The United Kingdom is pairing growth policy with workforce preparation. Its AI Opportunities Action Plan says AI should be treated as a major economic growth lever and states that Skills England will help prepare British workers for AI-powered industries. Government updates published in 2026 also report 1 million AI courses delivered since June 2025, as part of a wider commitment to upskill 10 million workers by 2030.
The United States is emphasizing workforce development and infrastructure. The White House’s 2025 AI education order says it is U.S. policy to promote AI literacy and proficiency by integrating AI into education, training educators, and developing an AI-ready workforce. Separately, America’s AI Action Plan includes a dedicated section on training a skilled workforce for AI infrastructure and says the U.S. must invest in the people who will build, operate, and maintain that infrastructure.
Canada is taking a more public-sector governance approach while also building broader workforce coordination. Canada’s new AI Strategy for the Federal Public Service says its vision is to serve Canadians better through responsible AI adoption, and its priority areas are framed around responsible use across government. Separately, Canada’s 2026 Workforce Alliances initiative says governments, employers, unions, educators, and Indigenous partners will work together to identify workforce pressures, map skills gaps, and coordinate sector workforce plans.
The policy gap governments are still trying to solve
The ILO says policy choices will strongly influence whether workers remain in transforming occupations and whether job quality rises or falls. It is urging governments, employers, and worker organizations to use social dialogue and proactive transition strategies.
The OECD is making a similar argument from a different angle. Its research says better worker outcomes are associated with training and worker consultation, while unmanaged AI deployment raises concerns about privacy, pressure to perform, and biased decisions. In finance and manufacturing, majorities of workers surveyed by the OECD said AI-related data collection increased performance pressure and raised privacy concerns.
That means governments are not only dealing with unemployment risk. They are also confronting questions about job quality, surveillance, bargaining power, retraining speed, and whether productivity gains from AI will be widely shared or concentrated. The current policy challenge is less “Will AI affect jobs?” and more “Who benefits from the transition, and how fast can labour-market institutions adapt?” This framing is an inference supported by OECD and ILO findings.
What appears likely next
The most defensible near-term view is that AI will replace some narrow job categories, shrink some teams, and automate many repeatable tasks, but the larger economic shift will be toward redesigned work rather than a one-time collapse in employment. Employers are investing in agents, automation layers, and AI infrastructure; governments are investing in compute, education, and reskilling; and labour institutions are trying to catch up.
The long-term goals are becoming clearer. Many companies want higher output with leaner, more AI-assisted teams. Many governments want national AI competitiveness without mass labour dislocation. Whether both goals can be achieved at the same time will depend on training systems, social protections, adoption speed, and how much control workers retain over AI-led workplace change. That conclusion is an inference drawn from the official policies and employer plans cited above.
Sources
International Labour Organization, OECD, World Economic Forum, European Commission, UK Government, White House, Government of Canada, Microsoft, Alphabet, Salesforce.
About the Author
Brad Socha is the founder of The Universal Record, focused on sourced, factual global reporting. Coverage includes international news, geopolitics, technology, and major developments.







