Microsoft said Friday that it would sell the military and intelligence agencies whatever advanced technologies they needed “to build a strong defense,” just months after Google told the Pentagon it would refuse to provide artificial intelligence products that could build more accurate drones or compete with China for next-generation weapons.
The announcement — made quietly in a small, town-hall-style meeting with the software giant’s leadership and employees on Thursday, then published on a blog Friday afternoon — starkly illustrated the radically different paths these leading U.S. technology companies are taking as they struggle with their role in creating a new generation of cyberweapons to help, and perhaps someday replace, American warriors.
But the divergent paths also underscore concerns inside the U.S. defense and intelligence establishments about how the United States will take on a rising China.
In the past two years, the Chinese government has set goals for dominance in the next decade in AI, quantum computing and other technologies it believes will allow its military and intelligence agencies to surpass those of the United States. Pentagon officials have questioned how committed domestic technology companies are to keeping the country on the leading edge, the way Raytheon, Boeing, IBM and McDonnell Douglas did during the Cold War.
Google encountered fierce opposition from young engineers to the company’s participation in Project Maven, a program to improve how drones recognize and select their targets. Google declared a few weeks ago it would not bid on a multibillion-dollar contract to provide the Pentagon with “cloud services” to store and process vast amounts of data. Amazon, for its part, appears willing to supply its services to the military and intelligence agencies, and it runs the information cloud services that power the CIA.
Even before Friday’s announcement, Microsoft seemed like the only plausible alternative for the Pentagon’s giant cloud project, called JEDI, in which Amazon is considered the front-runner.
But the announcement by Microsoft may have a greater effect on future technologies, including warning systems and weapons powered by AI. And the company’s leadership, after brief debates this summer, concluded that by dropping out of the bidding, Google was also losing any real influence in how the weapons would be used.
“This was not a hard decision,” Brad Smith, Microsoft’s president, said in an interview in his office. “Microsoft was born in the United States, is headquartered in the United States, and has grown up with all the benefits that have long come from being in this country.”
But Smith seemed to be trying to strike a middle ground.
He has sued the U.S. government repeatedly to halt Washington’s efforts to gain access to customer information stored in the company’s servers, and he is pressing for new international agreements to limit how the United States and its adversaries can use cyberweapons.
He also argued in the blog post that “to withdraw from this market is to reduce our opportunity to engage in the public debate about how new technologies can best be used in a responsible way.
“We are not going to withdraw from the future,” he said.
Smith’s comments stood in sharp contrast to statements by Google this month. When the company dropped out of the competition for the JEDI cloud computing contract, officials said they “couldn’t be assured that it would align with our AI principles,” a reference to a set of principles issued in June in which the company vowed not to design artificial intelligence products that would be used to harm people, or for surveillance or armaments. The contract is believed to amount to about $10 billion over a decade, but the shape of cloud computing that far into the future is hard to predict.
Some industry experts saw Google’s statement as an effort to gain political credit for backing out of a competition it was unlikely to win. Unlike Amazon and Microsoft, Google is missing some of the government certifications it would need to provide the software to the military.
Microsoft is believed to have a good shot, in part because President Donald Trump has made no secret of his distaste for Jeff Bezos, Amazon’s founder and the owner of The Washington Post. But even as Trump frequently dismisses the “Amazon Washington Post,” Amazon is considered the front-runner for the contract, in large part because of the company’s experience in building the CIA’s cloud computing ability over the last five years. In an analysis published in June, Deutsche Bank Research concluded that Amazon “is best positioned to win the lion’s share of the JEDI contract” and that Microsoft was the “main challenger.”
Culturally, Google and Microsoft are far different; while some Microsoft employees have expressed unhappiness with the company’s government contracts, most recently with Immigration and Customs Enforcement, the protests among Google engineers and other employees were far more intense and far more public.
Microsoft has been far more willing, however, to challenge the government. It went to court to try to keep the United States from being able to subpoena Microsoft for the emails or other records of customers whose data was stored abroad. (The suit was rendered moot by federal legislation.) It has been pressing for a “digital Geneva Convention” that would, like the traditional Geneva Convention, wall off certain civilian targets. The U.S. government has resisted the idea so far, not wanting to limit a president’s options to use cyberweapons against power grids or other targets on which civilians depend.
“We can’t control how the U.S. military uses our technology once we give it to them,” Smith argued. “But the military is subject to civilian control. And we believe we will have an opportunity to influence those discussions — but it’s not up to us.”
But to a large degree, the import of Microsoft’s announcement has more to do with the politics of confronting China and less to do with its immediate prospects. The Pentagon has struggled in recent years to bring the Silicon Valley innovation to new weapons systems and sensors. The acquisition and testing process is so long that many high-tech systems are outdated by the time they begin. While the Obama administration succeeded in creating an outpost in Silicon Valley, called DIUx, or Defense Innovation Unit Experimental, its budget was small and some promising off-the-shelf technologies — for example, small, inexpensive satellites that could be fired in large numbers over North Korea to detect mobile launchers — have been delayed for years.
But the next challenge facing both the government and the companies is whether U.S.-based companies will be limited in what kind of AI technologies they can sell to China.
In his interview, Smith would not indicate whether Microsoft would also provide all its products to, say, the People’s Liberation Army. “It’s an issue we are going to have to work through,” he said. But the reality, he and other executives in the technology field have noted, is that the Chinese would not accept U.S.-made artificial intelligence code, for fear that the United States would turn off Chinese access to the products in a time of conflict — or phone back to the Pentagon.
wraltechwire.com