It sounds like something out of the newest season of Westworld or Black Mirror: machines that can think and learn like humans. Yet every day, the world moves a little closer to a future propelled by the use of artificial intelligence (AI). Worries abound—from concerns about rogue killer robots to labor disruptions—but so does hope that AI could eliminate the need for menial tasks and launch a new era of prosperity. Which future is more likely, and what should policymakers do?
It is no secret that AI promises huge economic benefits. Even conservative estimates project at least $1.49 trillion in growth due to AI over ten years. Studies that focus on more specific applications of AI reveal even more potential. For example, AI-directed driverless cars could save the US about $1.3 trillion in annual costs per year, according to analysts from Morgan Stanley. While it is easy to get lost in the scale of these numbers, it is important to note that these statistics represent real quality of life improvements for working individuals and families across the country.
The benefits won’t just be economic, either. Other applications range from translating sign language to weather forecasting. AI has powerful applications in the medical field, where companies such as IBM have begun training AI to diagnose medical conditions. The financial savings from driverless car technology partially come from reduced traffic accidents, injuries, and fatalities. These opportunities should persuade policymakers that embracing AI will lead to more than convenience and economic growth. It will literally save lives.
Unfortunately, discussion of AI often centers on its drawbacks, real or imagined. While the public’s reservations about AI are often based more in fiction than scientific reality, even some expert commentators have fallen prey to more sophisticated myths.
One of the most egregious examples is the long-running fear that technological improvement destroys jobs. To be sure, there is always a period of friction during economic transitions. But in spite of many predictions, employment has not plummeted in the long run as a consequence of technological advancement.
In fact, in the long-term, technological development has improved working conditions and reduced the number of hours that employees must work—without triggering mass unemployment. Well-considered policies can reduce the transition costs for workers whose jobs will be replaced with AI. For example, Bank of America recently launched a program to train its employees whose jobs will be automated for other roles within the company. Public efforts should focus on helping these people go on to have productive careers rather than stopping technological advancement.
Another concern focuses on the privacy and security questions posed by AI. Especially when combined with huge data sets generated by online activity, artificial intelligence could lead to a detrimental imbalance of information as average Internet users become well known to the AI while knowing little or nothing about the AI or its observations. Many technology companies have already taken action to increase privacy and transparency, but some wonder if government should pre-empt these concerns with regulation.
Precautionary measures of this kind, while well-intentioned, are likely to do more harm than good. Especially considering the nascent nature of AI technology, policymakers are ill-positioned to project where and how AI might threaten privacy. This would force them to take haphazard or overbroad stances that could stymie important discoveries without fixing the privacy problem.
Instead, policymakers and tech leaders should address AI issues with a presumption of permissionless innovation. Entrepreneurs and researchers should be allowed to discover how AI can be most useful to society, and markets should determine what AI products are available to consumers. Rather than striking blindly at possible risks, policymakers should wait to act until well-defined problems present themselves. It is possible for governments to watchfully prohibit malicious practices without preemptively stalling AI’s many positive contributions. If leaders in this field are truly concerned about the future, this is the course they ought to chart.
Learn more at permissionlessinnovation.org
Keep reading: Artificial Intelligence and Public Policy
Photo credit: Mark Schiefelbein/AP/REX/Shutterstock