WASHINGTON, DC - OCTOBER 1: Arati Prabhakar, Director, Defense Advanced Research Projects Agency (DARPA) is interviewed by Washington Post Live editor Mary Jordan. (Kate Patterson/For The Washington Post)
Opinion writer

The federal government should take a lesson from DARPA, the Pentagon’s high-tech incubator. The agency behaves more like a Silicon Valley start-up than a bureaucracy. It takes risks that might fail, explores dark and potentially dangerous technologies and encourages a contrarian debate about science and the future. I wish more government agencies were as creative.

The Defense Advanced Research Projects Agency, as it’s formally known, is an outlier in a Washington colossus that generally seeks the safe middle zone. Many government employees these days seem terrified of making mistakes that might draw the wrath of the media or Congress. Too often, the government operates on a template that is low-risk and low-return. That’s true even at agencies such as the CIA and the State Department, which are paid to think outside the box.

Arati Prabhakar, DARPA’s director, told me in a recent interview that she tries to avoid worrying how a potentially controversial research project might look if it were bannered on the front page of The Washington Post. “One of our policies is not to back off these powerful technologies,” she explains. “If we’re cowed into not doing things because they might look funny, we won’t do our mission.”

Government use of “big data” is an area where Prabhakar thinks research should be pursued, accompanied by a policy debate about what rules are appropriate. As an example, she cites a DARPA program called Memex to develop search technologies that could be used to explore potential threats.

DARPA said last year that the Memex mission was to search online forums, chat groups, advertisements, job postings and hidden services to expose illegal sex trafficking. But Prabhakar disclosed last month that Memex is now helping track online activities of extremists linked with the Islamic State. This is a highly valuable tool, but it raises potential privacy issues, too.

“Virtually every area of technology has a double-edged sword,” Prabhakar argues. “Part of working on powerful technologies is to have the conversations” about how they should be used and consider the “potential for misuse by bad actors.” But those are decisions for policymakers, not technologists, she ­insists.

What makes DARPA unusual in the federal government is that it doesn’t back away from potential controversy or embarrassment. The agency was created in 1958 in the national humiliation that followed the Soviet Union’s successful launch of Sputnik into space. Its most famous success was the defense communications network originally called Arpanet, which became the backbone of the Internet. DARPA also helped develop stealth technology, the M-16 rifle and other innovative systems.

DARPA’s Web site lists dozens of innovative projects, from “unhackable” software to deep analysis of every aspect of social media. One project is dubbed “Automatic Detection of Psychological Distress Indicators in Online Forum Posts.” Others seek new technologies to detect and prevent “insider threats,” such as that posed by former National Security Agency consultant Edward Snowden.

To mobilize hackers to combat hacking, DARPA is holding what it calls the Cyber Grand Challenge. A news release describes it as “a competition that seeks to create automatic defensive systems capable of reasoning about flaws, formulating patches and deploying them on a network in real time.” The aim is to “overturn today’s attacker-dominated status quo.”

Prabhakar’s shorthand for this cyber research is that it’s an alternative to the current approach of “patch and pray.”

An applied physicist whose family emigrated from India when she was 3, Prabhakar got her start at DARPA and subsequently worked for a venture capital firm before returning to the agency. During the interview, she expressed some iconoclastic views about technology.

Prabhakar argued, for example, that Moore’s Law, which propounds that information storage and processing will become ever cheaper, “is slowing and could end.” Techno-optimists who assume ever-increasing complexity may have it wrong, she implied. The military should “rethink” complexity. “We’re at a point of diminishing returns in building ever bigger fighters, spacecraft” and other systems, she said.

For a military that has become addicted to technology, her comment is like taking away the cookie jar. Instead of increasingly complex systems, she says, the military should consider smaller satellites, cheaper launch vehicles and alternatives to fancy GPS-linked technologies that can be attacked. With new forms of inertial navigation to replace GPS, for example, “the future will be going to the past.”

Too often in conversations with U.S. officials, you hear explanations about why government can’t do things. This risk-averse culture limits the government’s ability to respond aggressively to challenges. We could use more of the brainpower and frankness that Prabhakar projects.

Read more from David Ignatius’s archive, follow him on Twitter or subscribe to his updates on Facebook.