Ctrl + N
The Pentagon is committing to prioritize ethics and safety as it ramps up its use of artificial intelligence.
It’s a pledge that could ease some of the concerns tech workers at Silicon Valley companies have raised about building powerful AI tools for warfare.
As the Pentagon on Tuesday laid out its first AI strategy yesterday to counter a surge of investment in the technology by China and Russia, it promised to work with industry leaders to “advance AI ethics and safety in the military context.” The Pentagon also promised to “articulate its vision and guiding principles for using AI in a lawful and ethical manner to promote our values.”
The decision to make this a core part of its strategy indicates the Pentagon has taken notice of the warnings that top technologists like Elon Musk have made for years about the dangers of building AI weapons that humans can’t fully control — and Google’s high-profile decision to pull out of a drone software contract with the military last year after employees raised ethical concerns.
Researchers concerned about the threat of AI weapons are welcoming the commitment — but with caution.
“It’s encouraging that the Pentagon’s says it will articulate its vision for using AI ethically and safely, but the crux is what they’ll actually articulate,” said Max Tegmark, a Massachusetts Institute of Technology professor and president of the Future of Life Institute, which conducts AI safety research.
“If they come out and say, ‘We’ll never build lethal autonomous weapons without meaningful human control,’ they’ll have an easier time recruiting top U.S. AI researchers than if they merely say ‘that’s classified.’”
The Pentagon is under pressure to be more transparent as it increasingly courts top AI researchers in the technology industry and academic community. The Google decision put a fine point on the broad gulf between the military and Silicon Valley — and it’s become a matter of national security for the Pentagon to bridge that divide. Experts warn a 21st century “arms race” against China to lead the world in AI is underway, and the government will need cooperation from top private sector talent to maintain its competitive edge.
“China has a clear and open strategy to be the world leader in AI,” said Peter W. Singer, a strategist at the New America think tank. “They also are creating and enforcing the kind of corporate partnerships between business and government that frankly aren’t possible here.” China exerts more influence over how much top technology companies like Baidu or Huawei cooperate with the government on its AI strategy than the U.S. does over businesses like Google.
The Pentagon’s strategy indicates that the Defense Department plans to do more to foster a better relationship with the tech industry. The Pentagon will focus on enhancing its partnerships with industry, seeking “bold new AI initiatives with large industrial partners, small start-ups, and venture capital firms,” the report said. The Pentagon also said it would remove crucial administrative hurdles to make it easier for industry to engage with the military.
Engaging with Silicon Valley on ethics and safety could be key to the success of any new AI initiatives. In its strategy, the Pentagon committed to invest in AI systems that minimize bias —a key concern industry is also grappling with as it invests heavily in the technology. And the Defense Innovation Board, the Pentagon’s key liaison to the technology industry, previously said it would develop AI ethics recommendations for the Defense Department this year.
The Pentagon released its strategy just a day after President Trump signed an executive order that aims to accelerate U.S. leadership in developing AI, largely through prioritizing AI research and development within the government. China has a had coordinated national strategy on AI for years, and the executive order and the Pentagon’s strategy appear to be the U.S. response.
The Pentagon won’t be the only agency that has to consider ethical questions as the Trump administration calls for all government agencies to prioritize AI. Given the high-stakes risks involved in how AI should be used by the military, the Pentagon may be forced to stay “a step ahead of the game” and engage in critical discussions about ethics with the tech industry early, said Jennifer Pahlka, the executive director of the nonprofit Code for America, which aims to bring digital services to government. She wants to see more open dialogue between the government and the technology industry about the tough questions that arise as AI is increasingly adopted.
“It’s critical to the future of the country that folks with very different perspectives talk more openly,” she told me.
BITS: T-Mobile and Sprint executives are testifying on the Hill today to defend their proposed merger, my colleague Brian Fung writes this morning. At the hearing, executives will face a grilling on whether the deal “will hurt competition, raise prices and harm economically struggling Americans,” Brian writes.
“The hearing by a key congressional panel also marks the first opportunity for lawmakers to personally confront John Legere, T-Mobile’s chief executive, over revelations that he and other company executives have been frequent guests at President Trump’s hotel in the District,” Brian said.
The hearing in front of the House Committee on Energy and Commerce comes as Sen. Richard Blumenthal (D-Conn.) and eight other Senate Democrats said they want the Federal Communications Commission and the Justice Department to reject the merger between T-Mobile and Sprint. In separate letters to FCC Chairman Ajit Pai and Makan Delrahim, the head of the Justice Department’s Antitrust Division, the senators said the merger would probably “raise prices for consumers, harm workers, stifle competition, exacerbate the digital divide, and undermine innovation.”
“This merger will turn the clock back, returning Americans to the dark days of heavily consolidated markets and less competition, with all of the resulting harms,” the senators said. “Our enforcement officials are the last line of defense preventing reconsolidation of our telecommunications markets at the expense of American consumers.”
NIBBLES: Health experts say Facebook must help curb the spread of disinformation from groups that question the validity of vaccines, the Guardian’s Ed Pilkington and Jessica Glenza reported. Facebook users have to be approved before they can join those groups on the social network — one of them, called Stop Mandatory Vaccination, has more than 150,000 authorized members. Some groups also claim that options such as taking high doses of vitamin C are a safer alternative to vaccines. The spread of such anti-vaccination messaging comes as Washington state last month declared a state of emergency due to a measles outbreak, Pilkington and Glenza noted.
“Parents deserve the truth,” Wendy Sue Swanson, who is spokeswoman for the American Academy of Pediatrics and has met with Facebook to bring up her worries about anti-vaccination messages on the platform, told the Guardian. “If they are being served up something that is not true it will likely increase their levels of anxiety and fear and potentially change their uptake of vaccines, which is dangerous.”
BYTES: Some major news publishers are balking at the financial terms of a news subscription service described as a “Netflix for news” that Apple plans to launch this year, the Wall Street Journal’s Benjamin Mullin, Lukas I. Alpert and Tripp Mickle reported. The Washington Post and the New York Times haven’t yet agreed to be part of the service partly because of the terms while the Journal has had productive exchanges with the tech company but has concerns too, per the Journal. Apple would plan to keep roughly half of the subscription revenue from that service — which would be incorporated as a paid option in the Apple News app — and publishers probably would not be able to access data about subscribers.
“Digital subscriptions are powering growth at big publishers including the Times, whose basic monthly subscription costs $15, the Post, which charges $10, and the Journal, which charges $39,” the Journal reported. “Some of those companies are skeptical about giving up too much control to Apple or cannibalizing their existing subscriptions to sign up lower-revenue Apple users, according to people familiar with the matter.”
— T-Mobile and Sprint were not among the more than 100 business leaders that urged Congress to pass legislation to protect immigrants brought to the United States illegally as children and known as “dreamers,” the Verge’s Makena Kelly reported. The heads of major tech companies such as Facebook, Twitter, Apple, Google and Amazon signed a letter to congressional leaders that was released on Monday. (Amazon founder and chief executive Jeffrey P. Bezos owns The Washington Post.)
“T-Mobile and Sprint did not respond to requests for comment, but their absence from the letter may have something to do with the proposed merger between the two companies,” according to the Verge.
— Microsoft co-founder Bill Gates said marginal tax rates in the United States could be higher but added that going too far could push people to move their money abroad, the Verge’s Nilay Patel reported. “I believe US tax rates can be more progressive,” Gates said, according to a transcript of his interview on the Vergecast podcast. “Now, you finally have some politicians who are so extreme that I’d say, ‘No, that’s even beyond.’ You do start to create tax dodging and disincentives, and an incentive to have the income show up in other countries and things.”
— More technology news from the private sector:
— The federal government will require owners of civilian drones to place the device’s registration number on the outside of the drone amid security concerns from law enforcement agencies, Bloomberg News’s Alan Levin reported. Some agencies worry that explosives could be hidden inside drones. “The regulation, which was posted Tuesday on a preview website for the Federal Register and takes effect Feb. 23, is part of an effort to bring more oversight to the rapidly growing hobby and commercial drone industry,” Bloomberg News reported.
— A poll found that a majority of voters in New York state and New York City support a deal for Amazon to open part of its second headquarters in the Queens neighborhood of Long Island City, the Wall Street Journal’s Jimmy Vielkind reported. The poll, which was conducted by the Siena College Research Institute, found that 56 percent of respondents supported the deal. “Among city residents polled by the upstate college, 58% support the project and 35% are opposed when told of both the promised jobs and incentives,” according to the Journal.
— Trump is expected to sign an executive order by Friday that would aim to protect U.S. telecommunications networks and could lead to bar Chinese companies including Huawei, The Washington Post’s Ellen Nakashima and Tony Romm reported. With the executive order, the president is expected to declare a national emergency about the threat to the supply chain. But the order won’t ban specific firms or countries, according to officials. “Major telecom companies such as ATT and Verizon already bar Huawei equipment from their core networks, a response to concerns raised years ago by U.S. intelligence agencies,” my colleagues reported. “But officials say that issuing the executive order now is a way to show the world that the United States is leading by example, taking decisive measures to protect the telecom supply chain.”
— More technology news from the public sector:
— Tech news generating buzz around the Web:
— Michelle DeBella, former global head of internal audit at Uber, is joining Lyft as vice president of finance strategy, TechCrunch’s Megan Rose Dickey reported.
— News about tech incidents and blunders:
— Today in funding news:
- House Energy and Commerce subcommittee on communications and technology hearing on the T-Mobile and Sprint merger.
- Senate Commerce Committee hearing on “policy principles for a federal data privacy framework” on Feb. 27.
Snow, extreme weather slam Hawaii:
Hundreds protest for immigrants with temporary status:
Do McDonald’s workers really only earn $7,000 a year? | Fact Checker: