According to external sources, Bill Gates recently proposed two more specific AI governance scenarios, one for taxing replacement machines and the other for setting “human reserves” for some jobs to mitigate the impact of automation on employment.
The robot tax is used to slow the replacement speed.
According to the article, Gates argued that the current tax system had an incentive effect on businesses to replace artificially with machines. Enterprises are required to pay the relevant taxes for their wages when they hire their employees, but the purchase of robots can usually be used as a more rapid credit for operating expenses, which will facilitate a faster shift to automation.
According to his vision, the taxation of such replacement machines could, on the one hand, slow down the pace at which enterprises cut down on manual jobs and, on the other hand, provide financing for vocational retraining and a stronger social security system.
Some jobs can limit AI intervention
Gates also suggested that part of the work could be classified as “human reservation”, i.e. limiting the use of AI in a given task. According to the article, this approach is relatively easy to promote at the policy level, as it can land directly through industry rules or job requirements.
He gave two main reasons. The first is the issue of job placement. The rapid taking over of a post by a machine may have a greater impact on people who have long been in a single industry and are more difficult to transfer. The second is the importance of human contact per se, especially in settings such as medical care, where certain messages are not suitable for direct machine delivery.
Rule-making remains to be clarified
The article mentions that Gates as a whole still belongs to the “Responsible AI” camp. He agreed with some AI practitioners calling for a slower advance of the front-line model, but also questioned whether the slowdown could be sustained in the long term. At the same time, he still sees the active role of AI in scientific research and medical health.
However, both scenarios still face real problems if they are to move forward, including who sets the rules, how the scope of application is to be defined and how the enforcement standard is to land. It was also mentioned that such measures could reduce the profit space of large AI laboratories.