Artificial intelligence is often portrayed as a game-changing technology capable of automating industries and replacing human workers; however, this narrative overlooks the significant human input behind AI systems, often involving low-paid, precarious work, writes researchers of the International Labour Organisation (ILO).

Uma Rani, Senior Economist at the organisation and Rishabh Kumar Dhir, Technical Officer at the organisation wrote that many AI systems rely on a “human-in-the-loop” model, where workers are essential for tasks like data annotation and system fine-tuning.

“An ILO survey of crowdworkers reveals that many of them are highly educated, holding bachelor’s or postgraduate degrees, often in specialized fields like science, technology, engineering, and mathematics (STEM). Yet such workers are primarily employed in routine and repetitive data work, which often require minimal specialized knowledge. This leads to a significant mismatch between the educational attainment of workers and the AI-related tasks that they are hired to do,” they wrote.

This invisible workforce is crucial to the development of AI, from self-driving cars to virtual assistants. Data for these systems must be meticulously labelled and categorised, tasks that require human intelligence.

Often outsourced to crowdworkers via digital platforms or AI-Business Process Outsourcing (AI-BPO) companies, these jobs are broken into microtasks, with workers paid small sums for each one. Despite AI being marketed as autonomous, it often relies on humans to bridge technological gaps, whether by transcribing audio, verifying outputs, or completing tasks AI struggles with.

“The reliance on human labour in the AI industry raises several critical questions. Firstly, there is a risk of deskilling workers and hindering their career development. Second, the shift towards contingent work arrangements can contribute to lower labour share in income, and it can increase income inequality,” they explained.

Even sophisticated AI systems, like large language models, depend on human trainers to improve responses and reduce harmful biases. This often exposes workers to distressing content, such as hate speech and graphic violence, which can have severe mental health consequences, including post-traumatic stress disorder and depression.

The reliance on human labour in AI raises critical questions about the future of work. Deskilling, career stagnation, and increased income inequality are significant risks, the researchers explained. While AI has the potential to boost productivity and create new jobs, its current structure highlights the urgent need for ethical standards and transparency in the industry, they added.

Governments and organisations are starting to address these challenges. Initiatives like the European Union’s AI Act and the Fairwork AI Principles aim to improve conditions for workers. Some platforms have adopted codes of conduct to protect workers’ rights, and certain companies are introducing wellbeing services for those exposed to harmful content.

“The future of work in the age of AI should be one of genuine collaboration between humans and machines, not one built on a hidden, global workforce facing decent work deficits. Only then can we realize the full potential of AI while ensuring a more equitable and sustainable future for all,” they added.