48% of workers fear losing their jobs to AI despite 90% surge in agent usage
A GMB Union survey reveals nearly half of employees feel their jobs are threatened as companies accelerate AI agent adoption.
June 14, 2026 · 3 min read
TL;DR: 48% of workers fear AI will end their jobs, just as AI agent usage surges 90% in the past year. The paradox reveals a gap between social fear and rapid business adoption.
What happened?
A report by GMB Union, a British trade union representing over 600,000 workers, reveals that 48% of respondents fear artificial intelligence will replace their jobs. The study, conducted in March 2025, surveyed 2,000 workers across various sectors, from finance to logistics. This figure comes amid a context where the use of AI agents — autonomous systems that perform tasks without human supervision — has grown by 90% in the past year, according to data from the same organization. The paradox is clear: while companies integrate AI at a breakneck pace, the workforce feels an existential threat. This is not the first time GMB Union has warned about automation; in 2017, it already cautioned that 4 million British jobs could be replaced by robots in two decades, according to a PwC study. Now, with generative AI, fear has accelerated: 72% of respondents believe their jobs will change significantly in the next three years.
Why is this important?
This study is not an isolated survey. It reflects a global trend: fear of automation has intensified as generative AI and autonomous agents demonstrate ever-wider capabilities. The difference from previous automation cycles — such as industrial robotization — is that white-collar workers now also feel vulnerable. A 2023 Goldman Sachs report estimated that AI could expose 300 million jobs to partial automation. The 90% increase in agent usage indicates that companies are moving from experimentation to mass deployment. Companies like Microsoft, Salesforce, and SAP already offer AI agents for administrative tasks, customer service, and data analysis. However, the fear is not just about losing jobs but also about losing control: 61% of respondents said they feel less autonomous in their work due to AI. Additionally, 55% believe AI benefits the company more than employees. This accelerates the need for reskilling and just transition policies, such as those proposed by the European Union with its AI Act, which requires transparency and training for affected workers.
Consequences and what readers should know
For workers, the recommendation is clear: invest in skills complementary to AI (critical thinking, creativity, team management) and stay updated on AI tools in their field. A 2024 McKinsey study suggests that demand for technical AI skills will grow by 55% by 2030, while social skills will increase by 30%. For companies, ignoring employee fear can lead to resistance, turnover, and talent loss. A notable case: in 2024, software company Intuit faced internal protests after announcing the replacement of 1,000 jobs with AI, leading to a drop in morale and several executives resigning. Governments should consider job retraining programs and social safety nets, such as France's plan allocating €2 billion to AI training by 2027. The future is not about total replacement but role redefinition: AI will automate tasks, not entire jobs, but the transition will be painful without preparation. For example, in the legal sector, AI already drafts contracts, but lawyers focus on strategy; in healthcare, AI-assisted diagnoses do not replace doctors but improve accuracy. However, 30% of GMB Union respondents believe their company does not offer sufficient AI training, exacerbating anxiety.
“AI will not replace workers, but workers who know how to use AI will replace those who don’t,” summarizes an analyst from TheVortiq. The key lies in adaptation: according to a University of Oxford study, workers who collaborate with AI increase their productivity by an average of 14%. But without inclusive policies, the 48% fear could become a self-fulfilling prophecy, where resistance to AI stifles innovation and competitiveness. Companies already deploying AI agents must clearly communicate benefits and offer reskilling pathways. Amazon's example, launching an AI training program for 100,000 employees in 2024, shows a possible path. Ultimately, the GMB Union study is a wake-up call: fear is not unfounded, but the response should not be panic, but collective action.