Bill Gates Warns AI Could Bring ‘Negative Surprises’ Within a Year
During an interview on CNN’s State of the Union, Bill Gates warned that rapidly developing AI could result in negative surprises over the next year. He said this in the context of bad actors eliminating safeguards from advanced AI models. Gates further emphasized that the imminent threat did not come from AI getting out of control on its own, but from bad actors manipulating AI for ulterior motives. He further noted that waiting for the industry to regulate itself is a recipe for disaster.
The need for Enforced External Monitoring
Gates proposed that the US government must set up dedicated safety agencies with the mandate to audit AI systems independently. Moreover, the government must ensure that every advanced AI model possesses built-in, non-removable protective barriers. In addition, AI models should also possess comprehensive monitoring capabilities. It is essential for politicians and law enforcement to establish standards of compliance and safety monitoring.
There is a unique structural flaw in AI regulation. Gates noted that historically, dangerous technologies such as nuclear weapons were funded and understood by governments first. This involved an examination of potential risks. However, AI emerged entirely from the private sector. Policymakers are oblivious to the safety risks posed by AI. Gates further spoke about the irony of the current circumstances where tech CEOs want to slow down AI development, but governments lack an enforceable mechanism to do the same.
Gates has also expressed that establishing a global AI framework will be significantly tougher than Cold War-era nuclear negotiations. With nuclear weapons, logistics were trackable. With AI, open-source models can be distributed globally and operated from anywhere without a trace.
The Token & Robot Tax
To mitigate the impending economic crisis, Gates has proposed a fiscal intervention. He proposes that the government levy a tax directly on AI tokens. These tokens are core units of text, code and data processed by an AI model. Revenue generated from these taxes would be used to fund support programmes for workers displaced by AI. The objective is to equalize the cost of human labor versus digital labor.
Gates also proposed completely prohibiting specific career fields from AI integration. It is important that certain occupations are legally reserved for humans to protect the psychological fabric of society. Doing so would ensure that human empathy remains relevant. These roles should include childcare workers, judges, and jurors. He further explained that determining which roles remain “legally human” will become one of the biggest political debates in the late 2020s.
