Trump orders some sort of vague action in the AI arms race

Through an executive order, President Donald Trump launched the American AI Initiative, further underscoring the importance of a group of technologies that are reshaping everything from medical diagnoses to war-fighting. The administration didn’t give many specifics in the order published Monday evening or details about funding for its various elements such as efforts to increase artificial intelligence-related educational opportunities.

A key element of the order involves marshaling AI technologies to further “national security interests against strategic competitors and adversarial nations.” Indeed, countries like China and Russia are increasingly focused on developing their own AI capabilities. China, for instance, released its plan for AI supremacy in 2017. Trump’s executive order calls for officials to develop an “action plan” within 120 days on how to protect the United States’ AI “advantage.”

The order calls for boosting AI skills in the workforce—directing agencies that award educational grants, for instance, to apply them toward AI-related pursuits. It also calls for agencies to begin developing regulatory and non-regulatory approaches to oversee key technologies and industries.

In a Wired op-ed to announce the executive order, Michael Kratsios, a top technology advisor to Trump, acknowledged concern over how AI development could impact civil liberties, writing that the technology must respect “American values.” Kratsios largely struck an optimistic tone on AI.

“We must invest in the industries of the future—and few industries are more important than AI,” Kratsios wrote. “If we do, we can create autonomous cars, industrial robots, algorithms for disease diagnosis, and more. However, we must act now to ensure this innovation generates excitement, rather than uncertainty.”

But if recent events are any indication, the public will likely remain uneasy with many AI technologies and applications. Researchers have shown, for example, that facial recognition software, an AI-based technology, can show biases against women and minorities. On the defense side, may tech industry workers worry about how militaries might use AI to develop lethal autonomous weapons. Google employees forced the company to abandon a partnership with the Defense Department once the company announced it was helping the military use AI to analyze footage from drones.

As AI reshapes warfare, the dual-use technologies it comprises are already impacting day-to-day life for Americans and, of course, people around the world. Perhaps it’s telling that in Kratsios’s Wired piece, he compares the American AI Initiative to a 1960s commission that grew out of concerns of “increasing automation.” According to the Pew Research Center, most Americans think robots and computers will displace workers from jobs (65 percent). If that happens, few Americans think the result will be more good jobs (25 percent), and most think that inequality between the rich and poor will be worse (76 percent).

Of the 1960s-era National Commission on Technology, Automation, and Economic Progress, Kratsios wrote that “leaders foresaw a world where technology could lead to a new era of economic prosperity—but only if we met the challenge head on.” Judging from the Pew numbers, the Trump administration may have a tough sell convincing the public that a new epoch of AI-enabled prosperity is on the way.