IBM chief executive Arvind Krishna caused a stir last year when he told the US Congress that he would “sunset” his company’s facial recognition software — and said it had been abused by law enforcement to carry out mass surveillance.
Tech groups Microsoft and Amazon have taken a similar stand on the controversial software, after pressure from employees. They said they would refuse to sell it to police forces and asked Congress to regulate the emerging technology.
These moves came after employees at Google compelled the company to step back from its work on artificial intelligence for the Pentagon, in 2018.
Some observers interpret these developments as Big Tech distancing itself from the government bodies so instrumental to its rise.
“I see these battle lines being drawn, where companies are saying in their mission statements what [policies] they want to be aligned with,” says Miriam Vogel, chief executive of EqualAI, a non-profit focused on removing unconscious bias from artificial intelligence.
She describes Silicon Valley’s relationship with the government and military as “deeply interconnected and increasingly fraught”, as the biggest tech companies’ valuations rise into trillions of dollars and their products now touch the lives of citizens on a daily, even hourly, basis.
Vogel, a former White House official and associate deputy attorney-general, says highly paid tech employees are voting with their feet, pressuring companies to be more vocal about the values they stand for — often in opposition to the concerns of security officials.
In recent years, the most prominent stand-off was when Apple refused to help the FBI break into an iPhone linked to the 2015 terrorist attack and mass shooting in San Bernardino, California — a stark demonstration of the company’s power and distinct set of values.
“Apple was steadfast in its protection of the privacy rights of its users [while] law enforcement desperately wanted to be stopping crimes before they happened,” says Vogel. “Their needs were at an absolute dichotomy.”
But for all the headlines suggesting that Big Tech is shunning work with the military, major deals continue.
Earlier this year, Microsoft won a 10-year, $22bn contract to supply 120,000 close-combat US soldiers with augmented reality headsets. In 2019, it was awarded a $10bn cloud computing contract for the Pentagon that many assumed was going to Amazon, which had also been an enthusiastic bidder.
Brandon Tseng, co-founder of Shield AI — a start-up helping the Pentagon build unmanned systems for conflict zones — says that, for every example of a Google stepping back, there is a Microsoft stepping in. “It’s a myth that talented engineers don’t want to work with the military,” claims Tseng, a former Navy Seal. “We’re close to 200 employees now, doubling year-to-year, and there’s tons of inbound interest . . . By and large, you find an enthusiastic workforce interested in helping the government solve these problems.”
Shield AI is among the companies thriving thanks to their unabashed support for the defence sector. Others include Palantir, the big data group co-founded by Peter Thiel that is now worth $40bn, and Anduril, which builds tech for border surveillance.
And the success of these groups partly reflects how the US Department of Defense is adapting to the tech culture — all too aware that its rigid hierarchies and traditions are no match for tapping in to private companies’ software-first approach to innovation.
“What you’re seeing is a trend of the US government embracing individuals that have come up in Silicon Valley companies,” says Darron Makrokanis, senior vice-president D2iQ, a cloud management start-up that helped the Air Force shift to remote working during the pandemic.
America’s tech sector has long been intertwined with the military, ever since the US government became a huge spender on early semiconductors and other expensive equipment that lacked a commercial market. Historian Margaret O’Mara, a professor at the University of Washington, has written that “whether their employees realise it or not, today’s tech giants all contain some defence-industry DNA”.
“Silicon Valley traces its origins to the Department of Defense and the aerospace and defence industry,” explains Yll Bajraktari, executive director of the National Security Commission on Artificial Intelligence, an independent group formed to make recommendations to Congress and the president.
Bajraktari rattles off technologies in which the military was instrumental. Radar, GPS and stealth technology all emerged during the cold war, he points out, not to mention Arpanet — the network founded in the 1970s that became the foundations of the internet.
He says these decades were shaped by the government being the primary investor and the largest buyer of the tech. But, as the internet became a part of everyday life, Silicon Valley shifted its focus to consumer and business applications — markets that became far larger than the US government orders.
“Over the last decade, the military realised it needed to engage more closely with the high-tech firms of the Valley to be able to compete with tech-enabled competitors such as China and Russia,” Bajraktari says.
In 2015, the US DoD created the Defense Innovation Unit, a Silicon Valley outpost designed to accelerate the Pentagon’s adoption of the newest tech as it seeks to deepen the tech-military alliance.
In its first years, the DIU struggled to gain traction, says Tom Foldesi, its director of commercial products and partnerships. “The department, in all honesty, had a bad reputation in the Valley,” he admits.
But Foldesi says it was not ethics but bureaucracy that was the top concern. “We were seen as a really challenging customer,” he explains. “The processes were opaque; decision making was slow. For a start-up with 12-18 months of runway, cultivating [the DoD] as a customer was a non-starter.”
As recently as three years ago, the DIU would issue solicitations and felt lucky to get two or three bids. However, that has changed dramatically thanks to streamlined processes and a better grasp of the tech culture. “Most of our solicitations are drawing bids from more than 100 companies — in certain cases we have had over 150,” Foldesi says.
He adds that he cannot recall a single conversation where the ethical dilemmas of working with the military have come up. “I know of some high-profile concerns that were raised at companies that I can count on one hand, but I’ve not found them to be representational of the ecosystem at large.”
To some, then, the refusals of IBM, Amazon and Microsoft to work on facial recognition technology are unusual. According to Tseng, the more common sentiment among those in the technology sector is “I’d rather be working side by side with our government customers than standing on the sidelines on some big important problems.”