Bill Gates made a widely-reported argument that if robots are replacing workers, and workers pay taxes, then shouldn’t robots be taxed as well?
Since then, California has moved ahead with a campaign to launch a state "payroll" tax on robots, while South Korea has implemented what is being called the world’s first tax on robots. The debate continues.
While what counts as a robot is not clear-cut and still open to much discussion, that point is less vital to the crux of the issue than the question of their effects.
If bringing in AI allows you to fire hundreds, if not thousands of people, that decision has inevitable broader effects – on consumer spending, on local communities and their social vitality, or health and wellbeing outcomes, and on local and national government income and payments.
Unless redressed via taxes, the sole beneficiary of such investment is the company, and a lucky few already better paid individuals, like executives and professional specialists, whose jobs are less likely to be automated in the first place.
Of course, some commentators have already noted that robots have been on the rise for decades. In addition, these are said to take over simpler, more automatic tasks, not entire jobs.
However, the robots of old aren’t the AI ones driving concerns today. As stressed in the University of Oxford study The Future of Employment, it is no longer just blue collar employees whose jobs are under threat, but increasingly white collar as well.
This makes it a problem of both scale and scope, threatening to squeeze out even further the middle classes on which Western economies and democracies have depended.
Another objection is the familiar one of tax as an impediment to corporate innovation. For instance, Andrus Ansip, the European Commissioner leading the European Union's efforts towards a digital single market, was quoted in February as being against Gates’ proposals, as these would in effect be "taxing progress".
He went on to suggest that "a lot of people will lose their jobs, but our municipalities, national parliaments, the European Union and also entrepreneurs and business people need to provide skills to those people who will lose their jobs".
The obvious question, however, is who will pay for providing those skills? If the answer is national governments alone, the same governments currently in a race to the bottom with regard to lowering their corporate taxes and so tax incomes, this risks further tipping the balance of benefits in favour of large multinational corporations in particular.
While redistributing gains via more employee ownership schemes may well be a better solution, it is a less likely one, and depends on corporations implementing these as well.
In short, while it may be entirely right for corporate executives to make investment decisions driven by bottom line considerations, it is similarly evident that these have broader effects, likely further perpetuating income and wealth inequalities.
Getting the practicalities of such a tax component will of course be difficult. For instance, we cannot always get information as to how many employees are displaced.
We also have a question of how that additional income will be spent beyond just retraining, and how we can ensure that it does exactly that.
However, allowing practicalities to get in the way of basic principles and aims is to take the easy way out. As Milton Friedman himself argued, profit-seeking behaviour from companies is fine, as long as they play by the rules of the game, which means abiding equally by taxes set by governments.
A "robot tax", as a targeted component of a broader redistributive burden on increasingly AI-based companies, seems a reasonable way of addressing this balance now, rather than later – as does ensuring corporations don’t avoid corporate taxes in the first place.
Maja Korica is associate professor of organisation and human resource management at Warwick Business School.