AI Tax: Call to Tax Robots Reopens the Debate

In an article published back in March 2026, we considered the issue of whether an AI tax and a four-day working week could form part of the response to AI-related job displacement. At that stage, the debate was largely focused on future risk: what might happen if artificial intelligence replaced large numbers of jobs, reduced demand for labour and left workers without clear routes into new employment.

That debate has since moved on. The issue is no longer confined to whether AI may affect work in the future. It is whether the tax system already gives businesses a financial incentive to replace people with technology.

In a recent interview, James Reed, the chairman and chief executive of Reed, has argued that the country should “back humans” and “tax robots”. In his view, the tax system currently places too much of the burden on employing people, while allowing businesses to reduce employment costs by using AI, chatbots, automation and robotics.

The main point being made in the argument is that employers pay wages, employer National Insurance, pension contributions, holiday pay, sick pay and other employment-related costs. A machine, chatbot or AI system does not. If the tax system makes human labour increasingly expensive while technology becomes cheaper and more capable, businesses will have a stronger financial incentive to automate.

The Tax Burden on Human Work

The timing of this debate matters. From April 2025, employer National Insurance contributions increased from 13.8% to 15%. At the same time, the secondary threshold at which employers begin paying those contributions was reduced from £9,100 to £5,000.

For employers, that change increased the cost of taking on staff. For lower-paid and entry-level roles, the effect can be particularly significant because the reduced threshold brings more employment within the charge.

This is why some business leaders describe employer National Insurance as a tax on jobs. The more people a business employs, the greater the employment tax cost. By contrast, a business that automates part of its workforce may reduce not only wage costs, but also employer National Insurance, pension contributions and other obligations linked to employment.

That does not mean every business adopting AI is acting irresponsibly. Many employers use AI to improve productivity, reduce repetitive work and help employees do their jobs more effectively. The issue arises where AI is used as a substitute for human labour rather than as a support tool.

Why the Labour Market Impact Matters

The debate over an AI tax reflects a wider concern that AI may change the structure of the labour market, particularly where technology is used to replace tasks that were previously carried out by employees.

The greatest concern is at the entry-level end of the jobs market. Administrative work, customer support, junior research, basic drafting, data handling and other routine office tasks have traditionally given graduates and school-leavers their first step into employment. These roles help young workers build confidence, learn workplace habits, gain supervision and develop the practical judgement that cannot easily be taught in the classroom.

Those are also the kinds of tasks that AI systems are increasingly able to perform. If employers use AI to reduce the need for junior staff, the impact may not be limited to immediate job losses. It may also weaken the future skills pipeline, because fewer people will have the opportunity to gain early experience before moving into more senior roles.

That creates a broader risk for employers as well as workers. Businesses may benefit from short-term cost savings, but later find that they have fewer trained employees ready to progress into mid-level and specialist positions. Social mobility may also suffer if young people without strong networks or financial support find it harder to access the first rung of the labour market.

An AI tax is therefore not simply a question of how technology should be taxed. It is also a question of how society protects entry routes into work, funds retraining and ensures that productivity gains do not come at the expense of long-term opportunity.

What Would an AI Tax Try to Achieve?

An AI tax would be designed to address the imbalance between taxing human work and not taxing automated substitutes for work. In its simplest form, the argument is that if a business replaces workers with AI, some of the savings should contribute to the public finances.

That revenue could then be used to support retraining, apprenticeships, employment support, skills programmes or other measures designed to help workers adapt. It could also help replace part of the tax revenue lost where work previously performed by employees is transferred to automated systems.

The purpose would not necessarily be to stop innovation. A carefully designed AI tax would distinguish between technology that supports employees and technology that directly replaces them. That distinction matters. AI used to make workers more effective is different from AI used to remove workers from the business altogether.

However, drawing that line in practice would be difficult.

AI Tax: The Practical Problems

The biggest challenge in relation to an AI tax is definition. What counts as AI replacing human labour? A factory robot may be easy to identify, but modern automation is often software-based. A chatbot that handles customer enquiries, an AI tool that drafts documents, automated scheduling software, self-service technology and AI-assisted research tools may all reduce the need for human work.

A second problem is measurement. Would an AI tax apply only when actual redundancies occur? What if a business does not dismiss anyone but simply stops recruiting because AI has reduced the need for junior staff? What if AI replaces only part of a role rather than the whole role?

There is also the risk that an AI tax could discouraging investment. The UK needs productivity growth. If an AI tax is too broad or too blunt, it may penalise businesses that are using technology responsibly to improve services, remain competitive or support their existing workforce.

This is why an AI tax would need to be carefully targeted. A poorly designed tax could create uncertainty, reduce innovation and become difficult to administer. It could also encourage businesses to move work offshore rather than keep activity in the UK.

Why the AI Tax Debate Is Not Going Away

Despite those difficulties, the AI tax debate is unlikely to disappear. The existing tax system was built around a labour market in which most economic value was created by human work. If more value is created by technology, automation and AI systems, governments will eventually have to ask whether the tax base should change.

The choice may not be between taxing AI and doing nothing. Other options could include reducing taxes on employment, increasing taxes on profits, creating levies linked to automation, funding retraining from general taxation, or offering incentives for businesses that use AI to augment rather than replace workers.

The important point is that employment taxes and automation incentives cannot be considered in isolation. If government policy makes hiring people more expensive while AI becomes cheaper and more capable, businesses will respond to those incentives.

The Question Policymakers Cannot Avoid

The debate over an AI tax has become more immediate. It is no longer only about what might happen if AI replaces jobs in the future. It is about whether the current tax system already makes replacing people with technology more attractive than employing them.

That does not mean an AI tax is automatically the right answer. A badly designed tax could penalise innovation, discourage investment and create uncertainty for employers using technology responsibly. Businesses should not be punished simply for improving productivity or using AI to support employees.

However, doing nothing also carries risk. If employment becomes more expensive while automation becomes cheaper and more capable, the tax system may gradually favour machines over people. That could reduce entry-level opportunities, weaken the future skills pipeline and shift more of the cost of disruption onto workers and the public finances.

The real issue is therefore one of balance. The UK needs productivity and technological investment, but it also needs employment, training, tax revenue and routes into work for young people. Any future AI tax would need to distinguish between technology that supports human work and technology that directly replaces it.

An AI tax may not be the complete solution to AI-related job displacement. But the underlying question is becoming harder to avoid: if AI systems increasingly do work that was previously done by people, how should the costs and benefits of that transition be shared?