Foreign media: Bill Gates recently published a long article on his personal website stating that AI's transition period could be “one of the most turbulent periods in human history”. He again proposed a tax on AI usage and robotics, while at the same time advocating the retention of a portion of the work that had to be carried out by people to mitigate the impact of automation on employment and the tax system.
In his view, the current tax system had a clear bias towards enterprises. Enterprises are required to pay wage taxes when employing their employees, but the purchase of robots or automated systems can often be treated as an expense to the enterprise. Such differences make it easier for enterprises to choose machines rather than to continue hiring.
Gates claims to tax AI usage
Gates proposes a two-part approach: a tax on AI token, which is the unit of measure used to process text and data in large models; and a direct tax on robots. In his view, such tax use would be more direct than simply raising the business tax, which ultimately affected only profits, which in a highly competitive environment could be so thin.
He also pointed out that labour had been consistently contributing to the social security system through wage taxes, but automation equipment itself did not bear a similar burden. This means that when an enterprise replaces labour with machinery, it not only reduces the cost of labour, but also weakens the financial resources that were otherwise dependent on wage taxes.
The human retention cap is about 40%.
Another central idea in the article was to create a category of “human reserves”. Gates believes that even if AI is technically able to do some of the work, society can still decide to leave these jobs to humanity.
Typical examples given by him include child care and jury services. Education and health care are in the middle: in the future, there will still be people in the room, but more use will be made of AI to expand service coverage.
As he had envisaged, even with the most radical version, the ceiling for such posts was approximately 40 per cent. He also divides protection into two categories: long-term retention, such as care work, and transitional protection, for those who find it difficult to transfer to work at short notice, such as older construction workers.
Gates thinks this round is faster.
In his paper, Gates stressed that this round of technological changes was different from the past. The shift to office work in agriculture had taken several generations and had led to a large number of new occupations requiring human judgement; however, in his view, AI could directly replace cognitive work in the areas of law, hospitality, medicine, software and manufacturing within a decade.
He quoted a similar view. Dario Amodei, Chief Executive Officer of Anthropic, has previously stated that at most half of the junior white-collar posts may disappear in the next five years.
However, the available statistics do not yet fully confirm large-scale unemployment. According to a study released by Yale Budget Laboratories and the Brookings Institution in October last year, there are no signs of wide-scale substitution in federal employment data as of July. The new AI unemployment tracking tool in California also failed to detect a state-wide loss of jobs, but unemployment has increased in highly educated positions that are more affected by AI.
Gates stated that he was willing to take a reputational risk for his judgement and believed that the coming years would determine whether AI would eventually close the gap or increase inequality.
Additional information:The concept of a “robots tax” was publicly introduced by Gates as early as 2017, but was opposed by a number of economists who viewed it as essentially a productivity tax.