It’s time to rethink the authorized remedy of robots
A pandemic is raging with devastating consequences, and longstanding problems with racial prejudice and political polarization are coming to a head. Artificial intelligence (AI) can help us meet these challenges. However, the risks of AI have become increasingly apparent. The scholarship has highlighted cases of AI opacity and lack of explainability, design choices that lead to bias, negative effects on personal well-being and social interactions, and changes in the power dynamics between individuals, corporations, and the state that contribute to growing inequalities. Whether AI is developed and used in a good or harmful way depends to a large extent on the legal framework that governs and regulates it.
There should be a new principle for AI regulation, a principle of legal neutrality of AI that states that the law should tend not to distinguish between AI and human behavior. Currently the legal system is not neutral. An AI that is much safer than a person may be the best choice for driving a vehicle, but existing laws can prohibit driverless vehicles. A person can produce higher quality goods than a robot at a similar cost, but a company can automate because it saves taxes. AI is better at generating certain types of innovation, but companies may not want to use AI if it restricts ownership of intellectual property rights. In all of these cases, neutral legal treatment would ultimately benefit human wellbeing by helping the law better achieve its underlying policy goals.
CAMBRIDGE UNIVERSITY PRESS
Look at the American tax system. AI and humans engage in the same commercially productive activities – but the companies they work for are taxed differently depending on who or what is doing the job. For example, automation can help companies avoid wage taxes from employers. If a chatbot costs a company as much as it does before taxes as an employee doing the same job (or even a little more), automating the company costs less after taxes.
In addition to avoiding payroll taxes, companies can expedite tax deductions for some AI if they have a physical component or fall under certain software exemptions. In other words, employers can claim a large portion of the cost of some AI as a tax deduction upfront. Finally, employers also receive a number of indirect tax incentives for automation. In short, while tax laws weren't designed to encourage automation, they prefer AI over humans because labor is taxed more than capital.
And AI pays no taxes! Income and employment taxes are the state's largest sources of income and together account for nearly 90% of total federal tax revenues. Not only does AI pay no income taxes or generate labor taxes, it does not buy goods and services, it does not collect sales taxes, and it does not buy or own any property, so it does not pay property taxes. AI just isn't a taxpayer. If all work was automated tomorrow, most of the tax base would disappear instantly.
When companies automate, the government loses revenue, potentially hundreds of billions of dollars in total. This can severely limit the government's solvency for things like social security, national defense, and health care. If at some point people get comparable jobs, the loss of income is only temporary. However, if the loss of jobs is permanent, the entire tax structure will have to change.
The robot taxation debate began in 2017 after the European Parliament rejected a proposal to consider a robot tax and Bill Gates subsequently endorsed the idea of a tax. The problem is even more critical today as companies turn to the use of robots due to the risk of pandemics for workers. Many companies ask: why not replace people with machines?
Automation should generally not be discouraged. However, it is important to develop tax-neutral guidelines to avoid subsidizing inefficient technology uses and to secure government revenues. Automation for tax saving purposes may not make businesses more productive or consumer benefits, and it may lead to a loss of productivity to reduce the tax burden. This is not socially beneficial.
The benefit of tax neutrality between humans and AI is that the market can adapt without tax distortions. Companies should only automate if it is more efficient or productive. As the current tax system favors automation, a transition to a neutral tax system would make workers more attractive. Should the pessimistic prediction of a future with significantly higher unemployment due to automation prove correct, neutral tax revenues could be used to improve the education and training of workers and even to support welfare programs such as basic income.
Once policymakers agree that they do not want to use AI against human workers, they could lower taxes on humans or reduce the tax benefits for AI. For example, wage taxes (which companies levy on the salaries of their workers) should potentially be abolished, which would encourage neutrality, reduce tax complexity and end the taxation of something of social value – human labor.
More ambitiously, the legal neutrality of AI can lead to a more fundamental change in the taxation of capital. Although new tax systems could target AI directly, doing so would likely increase compliance costs and make the tax system more complex. It would also mean “tax innovation” in the sense that it could penalize business models that are legitimately more productive with less human labor. A better solution would be to increase capital gains taxes and corporate tax rates to reduce reliance on sources of income such as income and wage taxes. Even before AI hit the scene, some tax experts had argued for years that taxes on labor income were too high compared to other taxes. The AI could provide the necessary impetus to finally address this problem.
Opponents of increased capital taxation base their arguments largely on concerns about international competition. For example, Harvard economist Lawrence Summers argues that "taxes on technology are more likely to drive production offshore than create jobs on-site." These concerns are exaggerated, especially with countries like the United States. Investors are likely to continue investing even with relatively high taxes in the US for a variety of reasons: access to consumer and financial markets, a predictable and transparent legal system, and a well-developed workforce, infrastructure and technological environment.
A tax system shaped by the legal neutrality of AI would not only improve trade by removing inefficient subsidies for automation. This would help ensure that the benefits of AI do not come at the expense of those most at risk by creating a level playing field for human workers and ensuring adequate tax revenues. AI is likely to lead to massive but poorly distributed financial gains, and this requires and enables policymakers to rethink how they allocate resources and distribute wealth. You may find that we are not doing this well now.