The Transformation: Artificial intelligence has arrived in Washington, and it’s playing bigger than ever before – For years, the government approached artificial intelligence with the same mix of fascination and trepidation that many Americans had about robots replacing human workers.
In the past year, however, the Obama administration has embarked on a mission to make AI more mainstream – and important – to the nation’s capital than ever before.
That transformation is coming not from Silicon Valley – though tech executives such as Alphabet’s Eric Schmidt and Facebook’s Peter Thiel played roles – but from the Oval Office.
A presidential memo issued in April 2015, which gave federal agencies permission to “use AI technologies as a solution to real problems of national importance,” was a turning point.
Now nearly 3,000 “AI applications” – or instances of AI in use – exist across 29 federal agencies, compared with a handful of pilot projects a year ago, according to estimates.
These are not merely AI-enabled calculators in the Pentagon’s basement. AI is being used for everyday purposes such as facial recognition at border patrol, data mining at the FBI, and medical analysis at the Department of Veterans Affairs, among others.
Use of AI “has moved out of the research and development phase and into the operational phase,” said David Smoley, chief information officer at the Veterans Affairs Department, where AI is being used to analyze health records to identify patients who are at the highest risk of suicide.
The rationale: That this is not the beginning of the end, and the administration is not succumbing to the temptation to build Skynet, the fictional artificial intelligence system from the Terminator films that begins a nuclear war.
The White House frames AI as a “competitive advantage in the world economy,” akin to past economic disruptors such as the Industrial Revolution and the Internet.
The most recent report from the administration on the subject, released last week, referred to the spread of AI as an “economic transformation” that “presents a broad set of technological, economic, and societal opportunities and challenges.”
Concerns: Many AI experts and researchers have urged caution on the part of the government, saying that AI’s entry into the national life – especially areas as sensitive as national security and law enforcement – was not thought through.
A paper released by the Center for New American Security think tank last month noted that AI could increase government efficiency but could also “undermine public confidence” if systems fail or their decision making can’t be explained.
It also said that the government’s lag in AI adoption could leave the country unprepared for AI-enabled attacks, which could move faster than the traditional pace of government.
A 2015 report from the Pew Research Center warned that AI would displace more jobs than it created and that the economy would be broadly transformed, in ways that are difficult to predict.
Room for debate: So is this a good thing or a bad thing? While some people fear that AI systems will replace human workers or enable the government to snoop on them, many of the intended applications of AI are benevolent – to speed up routine government services, make sense of reams of data or predict national security threats.
The question is whether the government is moving too fast and with too little oversight, risking blows to the public’s confidence in AI systems if something goes wrong. [How the jobs market is already being changed by machines and artificial intelligence]
The AI push: AI has been on the radar of the White House since 2011, when Schmidt, then chief executive of Google, was part of a working group on science and technology.
That led to an initial report on AI in 2014 and then to the president’s memo in April 2015 calling for agencies to adopt AI.
Since then, the White House Office of Science and Technology Policy has led the charge in promoting AI throughout the government, aided by an industry group called the Information Technology Industry Council, whose members include companies such as Apple, Cisco, Google and Intel.
Last year, OSTP hosted five public workshops on AI to encourage adoption in areas such as law enforcement, national security and education.
Those sessions were followed by the establishment of an interagency task force, called the National Science and Technology Council Subcommittee on Machine Learning and Artificial Intelligence.
That task force, whose 12 members include representatives of the departments of Commerce, Defense and Energy and the National Science Foundation, has been meeting monthly to develop plans to integrate AI across government agencies.
The task force also recommended the creation of a new AI unit within the White House to coordinate AI work across agencies, according to a government official who spoke on the condition of anonymity.
Current and former White House officials say that the government’s embrace of AI is a reaction to the technology’s rapid commercialization and the prospect that it could be used for military purposes by other countries.
“It’s one of these situations where, if you don’t take a proactive approach, you might find yourself in a situation where you’re playing catch-up,” said John Holdren, the president’s science adviser.
What’s next? Although much of the work so far has focused on using AI for routine or analytical tasks, the government is already thinking about how AI systems could make decisions on their own – such as pulling the trigger on a weapon, in the case of the military – and how those systems would be designed to explain their actions after the fact.
“The agencies are starting to explore areas where AI could be used in a more autonomous way,” said Terah Lyons, the White House’s AI policy expert. “That could be in decision making but also in areas like health-care delivery.”
OSTP also released a report last week calling for greater research into AI’s societal implications and urging agencies to make public data available that could aid AI research.
For now, the spread of AI into everyday government operations seems likely to continue, especially with other countries such as the United Kingdom and Canada adopting similar plans.
“We recognize that we need to stay on top of this, to keep pace with it and to try to stay ahead of it as much as possible,” said Lisa Porter, the deputy undersecretary of defense for research and engineering, in a speech in July.
Read more: Donald Trump and the cult of personality in artificial intelligence Future Tense is a partnership of Slate, New America, and Arizona State University that examines emerging technologies, public policy, and society.