Her job: Ensuring AI and radical climate fixes don’t backfire

By Chelsea Harvey | 05/09/2024 06:30 AM EDT

As Biden’s top science official, Arati Prabhakar balances the promise and peril of new technologies.

Arati Prabhakar, then the nominee to serve as director of the White House Office of Science and Technology Policy, testifies July 20, 2022, before the Senate Commerce, Science, and Transportation Committee.

Arati Prabhakar, then the nominee to serve as director of the White House Office of Science and Technology Policy, testifies July 20, 2022, before the Senate Commerce, Science, and Transportation Committee. Francis Chung/POLITICO

Arati Prabhakar wrestles everyday with questions that could make or break the future of humankind.

What safeguards are appropriate for artificial intelligence? When is it OK to share someone’s genetic information? Should humans attempt to reverse global warming — even if the methods pose serious risks to the planet?

None of the questions has easy answers.


But for Prabhakar, who serves as the nation’s top science adviser, it’s critical that the people in charge continue to ask whether the promises of scientific advances are worth the peril.

“All of human history tells us that powerful technologies get used for good and for ill,” Prabhakar told E&E News in an interview. “Our job is to make sure that we manage the risks so that we can seize the benefits.”

Much of Prabhakar’s work occurs outside the public view. Yet few people have as much power to bend the president’s ear on how science and tech figures into federal policy. Prabhakar heads the White House Office of Science and Technology Policy, and she’s also the official assistant to the president for science and technology, giving her a direct line of influence to the Oval Office.

But her efforts can have major public ripples, especially when they concern cutting-edge technology.

Over the last two years, OSTP has helped craft new guidelines for artificial intelligence, including a presidential executive order and a long-awaited — but nonbinding — AI Bill of Rights, which outlined steps companies can take to safeguard the public from AI’s potential dangers. In March, the White House Office of Management and Budget issued a new policy requiring federal agencies to implement risk-management practices, publicly disclose information about their own AI use and designate chief AI officers.

OSTP also waded into the contentious debate on climate geoengineering last summer with its first report on solar radiation modification, a hypothetical — and highly controversial — proposal to reverse global warming and cool the planet by blocking the sun’s rays. The report was mandated by Congress, meaning it wasn’t a choice. But it still made a splash as the White House cautiously acknowledged, for the first time, that it would be open to studying the disputed technology.

Beyond those two hot-button issues, OSTP over the last few years has helped support President Joe Biden’s Cancer Moonshot, established a new health research agency, advanced low-carbon technologies and developed new strategies for monitoring greenhouse gas emissions.

“If we don’t take these issues seriously, if we don’t deal with the climate, if we don’t deal with AI, … that’s not a future I want to live in,” Prabhakar said.

‘Humans are the ones who have values’

Prabhakar has been breaking barriers — both social and scientific — for most of her life.

Born in India and raised in Chicago and Texas after immigrating to the U.S. at age 3, she was the first woman to receive a doctorate in applied physics from California Institute of Technology, where she also earned her master’s in electrical engineering.

She knew before she graduated that she didn’t want a traditional career in academia. So at the suggestion of her thesis adviser, she decided to accept a congressional fellowship with the Office of Technology Assessment.

The entire decision was made “purely on a lark,” she said in a 2020 oral history interview for the American Institute of Physics. “I had no view, negative or positive, about public service, about the role of the federal government,” she said. “And it completely changed everything. It was great.”

It was the beginning of a swift ascent. When the fellowship ended, Prabhakar went to work as a program manager for the U.S. Defense Advanced Research Projects Agency — the Pentagon’s most important research and development organization — focusing on semiconductor technology.

She eventually became the founding director of DARPA’s Microelectronics Technology Office. She continued to supervise semiconductor research there, as well as projects involving nanoelectronics, infrared imaging and optoelectronics, which are devices that detect or emit light.

Prabhakar made history again in 1993 as the first woman to lead the National Institute of Standards and Technology, unanimously confirmed by the Senate at age 34. At the time, the Clinton administration was working to expand federal efforts to support U.S. technology companies, with NIST leading the way.

“She is a very, very capable manager,” said Craig Fields, former head of DARPA and Prabhakar’s previous boss there, in a 1993 interview with the scientific journal Nature. “She’s innovative and she understands new technology and how to apply it.”

In 1997, Prabhakar left NIST for a 15-year stint in Silicon Valley, bouncing from the Raychem Corp. to technology incubator Interval Research and finally landing as a partner at the tech-focused investment firm U.S. Venture Partners, where she stayed for a decade.

But it was her next post that would most prepare her for the delicacies and dangers of emerging tech. In 2012, she returned to Washington as the head of DARPA under President Barack Obama — an agency tasked with developing advanced military technologies, including weapons designed to take human lives.

DARPA expanded its Squad X initiative during Prabhakar’s tenure, a program aimed at equipping infantry with more advanced battlefield technology such as handheld drones or “smart” target-tracking munitions.

President Biden meets with his Council of Advisors on Science and Technology.
Arati Prabhakar (second from left) has served as the president’s science adviser since October 2022, making her one of the nation’s most powerful influences on federal science and technology policy. | @POTUS/X

Prabhakar also oversaw the creation of the agency’s Biological Technologies Office, a division that immediately promised to raise new ethical questions about the use of biotechnology in national security. The office has explored everything from mental health interventions to neural implants and brain-controlled prosthetics, and it helped pioneer the mRNA technology later used to develop Covid-19 vaccines.

“I used to say to my folks at DARPA, ‘If your day job is creating powerful technologies, you are in the business of figuring out what the ethics of those technologies are,’” Prabhakar said. “The starting point for me with any powerful new technology is to recognize that it is value-neutral. Technology doesn’t have value — we humans are the ones who have values.”

For Prabhakar as Biden’s science adviser, AI became an immediate testing ground for that philosophy.

“When we started digging in on AI, the range of issues we had to contend with had to do with safety and security, but also the potential to exacerbate discrimination. There are issues of privacy and civil liberties, there are issues about jobs and employment,” Prabhakar said. “All of those had to be dealt with.”

Those conversations led to a 2023 executive order that created expanded guardrails for federal agencies and some AI development companies. Permanently enshrining those safeguards will require an act of Congress. But OSTP’s movement on AI is still some of her fondest work of the last year and a half.

“It was a place and a time when what we do at OSTP was critically important to the president and to the country, and it’s something I’m very proud of,” she said.

‘Bubbling, boiling issues’

Prabhakar joined OSTP in October 2022, replacing then-acting Director Alondra Nelson, who took over earlier that year after Eric Lander resigned amid reports of bullying. She immediately made history as the first woman, immigrant or person of color to be confirmed in that position by the Senate.

She was swiftly hailed by scientists and policy experts alike for her impressive resume and extensive experience in both industry and government.

“We all cheered when Arati was named,” said Jeannette Wing, executive vice president for research at Columbia University and a former assistant director for computer and information science and engineering at the National Science Foundation. “We all knew her, and so we thought it was a great choice.”

Prabhakar’s diverse experience made her an ideal candidate for a job that requires extensive coordination among public and private entities, said Harvard University scientist John Holdren, who served as Obama’s science adviser from 2009 to 2017.

“She’s very smart. She’s very focused. She’s very committed. She’s very thoughtful. She’s very energetic. And she is, in my judgment, an effective leader,” he said. “She has a fabulous background for the position that she’s now in as the president’s science adviser.”

Despite a lengthy list of credentials, it’s not an easy time to be the nation’s top voice on science and tech.

It’s a historic moment for ethical conundrums in research and technology development. AI, virtual reality, gene editing, advanced biotechnologies, carbon removal, geoengineering and quantum computing are just a few of the rapidly evolving innovations raising tough questions about safety and equity. Some, like AI and geoengineering, could even pose existential threats to humanity if misused.

Meanwhile, world leaders have been forced to grapple with cascading global threats the last few years, including the Covid-19 pandemic, the accelerating impacts of climate change and escalating global conflicts — all of which have raised the need for both more scientific and technological innovation and more discussions on the ethics of deploying them.

These concerns have driven a cultural shift in public attitudes toward technology, said Suresh Venkatasubramanian, director of the Center for Technological Responsibility, Reimagination and Redesign at Brown University and formerly the assistant director for science and justice at OSTP.

“Three to four years ago, it felt like technology was viewed as a tool, like Microsoft Office: a thing we should use and use more of,” he said. “I think that’s changed. Technology is a thing we need policy action around.”

While Prabhakar has happily led the charge on matters like AI and cutting-edge cancer interventions, her office has been less enthusiastic to weigh in on at least one major issue of the times.

OSTP quietly dropped its report on solar geoengineering on a Friday afternoon last summer leading into Fourth of July holiday weekend — a move that industry experts have widely regarded as an effort to keep the document out of the spotlight. OSTP also has heavily emphasized the fact the report was required by Congress and not an independent initiative.

“In every administration, solar geoengineering has been a very controversial topic,” said Shuchi Talati, founder of the Alliance for Just Deliberation on Solar Geoengineering, a nonprofit that advocates for stronger governance over the controversial technology. “And as this administration is thinking through really massive climate policy, and the most climate actions any administration has ever taken, I think the notion of engaging solar geoengineering was challenging.”

Solar geoengineering is one of the most controversial subjects in the climate science community. In theory, world leaders could artificially cool the planet by spraying reflective aerosols into the atmosphere and beaming sunlight away from Earth. But researchers warn that doing so could result in a variety of poorly understood side effects, like damage to the planet’s ozone layer or disruptions to global precipitation patterns.

And once begun, solar geoengineering would be difficult to safely stop. Halting it too quickly could send global temperatures skyrocketing faster than Earth’s ecosystems could handle, a phenomenon scientists have grimly dubbed “termination shock.”

The OSTP report avoided a strong stance on the subject, cautiously suggesting that it would be open to further study of the technology without actually committing to a national research program. Still, the fact the White House formally addressed the question — even if forced to do so — immediately risked politicizing the issue and polarizing national conversations on the subject, Talati said.

“So I understand the reticence about it,” she added.

Prabhakar acknowledged it’s one of the trickier climate policy debates, with big questions surrounding the ethics of field experiments and the logistics of governing and regulating them.

“There are many issues about safety and the impacts this kind of research can have,” Prabhakar said. “On top of that, there are concerns that this can become a distraction from the work that absolutely has to be done to decarbonize our economies and build resilience.

It’s full of these bubbling, boiling issues.”

OSTP’s climate efforts instead have focused on the more immediate issue of decarbonization — striving to meet the Biden administration’s goal of net-zero emissions by the year 2050.

“There was an emphasis at OSTP on ensuring that we are delivering the speed and scale needed to get to net-zero emissions,” said Costa Samaras, director of Carnegie Mellon University’s Wilton E. Scott Institute for Energy Innovation and OSTP’s chief adviser for energy policy from 2021 until January of this year.

That still means plenty of work on new and developing technologies, including alternative fuels and fusion energy. But some of the OSTP’s most important climate work has been its most understated.

In November, for instance, the department released a new national strategy for measuring and monitoring greenhouse gas emissions — one of the most basic steps toward ensuring carbon emissions are falling fast enough to meet the administration’s goals.

Still, knottier questions around issues such as geoengineering may loom larger for the White House as global temperatures continue to rise.

Some research groups already have begun quietly conducting their own climate engineering experiments, and interest in the subject continues to climb. Establishing a federal research program may be useful to standardize and enforce rules for experiments, said Frank Keutsch, a scientist and geoengineering expert at Harvard.

“The need is that there’s some part in the federal system that can be very nimble and fast to respond to these emerging technologies,” he said.

Global challenges like climate change mean science and technology will play an even greater part in federal policy, added Samaras, the former OSTP energy adviser. Same goes for the president’s science adviser.

“I think that the days of OSTP being unknown are behind us, and that OSTP will be an important component of any White House going forward, given the accelerating role science and technology has in our lives,” he said.

Despite all the controversies, hazards and unknowns of her day-to-day work, Prabhakar says she’s still starry-eyed about the ways breakthrough technologies can make the world better.

“Science and technology actually gives you a way to live in a more hopeful future,” she said. “That’s actually our biggest job.”