Training workers in AI is one way to cauterise potential government revenue losses.
At a glance
- AI’s impact on jobs and tax revenue is still uncertain, sparking policy debate.
- While AI may shrink the tax base, it could also drive new government revenue opportunities.
- Proposals include reskilling workers, taxing AI usage and updating employment laws.
In August, UK bank workers’ union Accord urged banks, accounting and insurance firms to prepare to reskill millions of employees who could be replaced by AI. A month earlier, the Australian Adopting Artificial Intelligence inquiry heard calls from the Media, Entertainment and Arts Alliance to impose taxes on companies that replaced their workforce with AI.
Meanwhile, policy experts debate the risks and benefits of generative AI regarding wage and wealth inequality and job displacement, with various fiscal policy options being explored. Some theorise over eight million Britons could lose jobs, while others argue job augmentation is more likely and we could see enhanced productivity.
Impact on the tax base
This advancing technology could replace human labour in various sectors, with administrative and secretarial occupations most at risk. Naturally, concerns have been raised about rising unemployment and a shrinking tax base. But the long-term implications are still being debated.
Academics, including Professors Chand, Kostić and Reis, argue focusing solely on job losses as a reduction in tax is too narrow. Their research suggests the rise of AI is interwoven with broader economic, demographic, climate and social changes that need to be better accounted for in current and future tax policy. For example, AI is driving change in global migration and climate adaptation which could influence corporate and individual tax bases.
In the UK, Ben Lee, partner at Andersen LLP, believes AI won’t necessarily limit government revenue.
“AI also presents individuals and businesses with significant opportunities to generate wealth, expanding the overall tax base,” he says.
“Many of our clients have successfully developed profitable AI models utilising digital assets, so I question whether the impact of AI on government revenue will be as detrimental as some predict. There’s also the potential for AI to be used by tax authorities to streamline and enhance investigations, which could increase revenue.”
Policy and regulatory challenges
One thing most experts agree upon is that swift regulatory changes are needed. According to Lee, the focus should shift towards workforce considerations.
“Current regulations, like the EU’s AI Act, prioritise protecting user rights and safety, particularly around data sharing. Although crucial, we’re yet to see meaningful discussions or policy developments regarding employment rights and the implications of AI on livelihoods,” he says.
Various policy ideas are currently being proposed and debated in real time. These include taxing AI manufacturers, imposing levies on companies that replace human workers with AI on a ‘per use’ basis, and even increasing individual tax burdens.
The Institute for Public Policy Research (IPPR) argues a job-centric industrial strategy, including a variety of policy interventions, is required to limit job displacement and encourage significant economic gains.
Even OpenAI CEO Sam Altman called for a wealth tax to ensure the potentially large gains from the deployment of generative AI are equally shared.
Professors Chand, Kostić and Reis advocate for neutrality, simplicity, efficiency, fairness and flexibility in taxation policy. They argue some policy ideas, such as taxing AI as independent entities, would contravene these principles.
Lee agrees there are challenges in the practicalities of some proposed policies.
“It’s challenging to see how taxing businesses that replace human workers with AI would work in practice. Accurately attributing specific job losses to AI would be highly complex and such a policy could be prone to manipulation,” he says. “Additionally, it’s unclear if we will actually see a net decrease in employment, as many roles are likely to evolve to work alongside AI rather than be completely replaced.”
Workforce adaptation
Professors Chand, Kostić and Reis recommend implementing an ‘education tax’ to finance programmes that reskill workers displaced by automation. Pointing to similar successful education taxes in India, Nigeria and Jamaica, they propose both individuals and businesses contribute on a means-tested basis.
The IPPR also argues for policy that proactively creates jobs with low risk of automation and addresses our undersupply of vocations like social care and mental health services by offering to retrain workers. One idea put forward is for a National Employment Service to support people through labour transitions.
Lee believes safeguarding workforce stability is paramount and could be supported through employment law changes, as well as reskilling initiatives.
“As the workforce landscape inevitably shifts, reskilling workers who lose part or all of their roles to automation will be essential,” he says. “Revisiting redundancy mandates could be a strong first step in addressing these changes.”
While AI’s impact on the workforce and tax base is still uncertain, one thing seems clear: thoughtful policy is needed to balance innovation with revenue stability. Emerging technology may offer opportunities to expand the tax base and streamline tax compliance, but potential job displacement and augmentation is real and will require reskilling strategies.
The IFA international conference online 2024 will explore AI on 7 November. Register HERE.
