Later On

A blog written for those whose interests more or less match mine.

The Chinese threat that an aircraft carrier can’t stop

leave a comment »

UPDATE below.

If you like techno-science-fiction, you surely must read Daniel Suarez’s novel Kill Decision, which pretty much lays out the issues underlying David Ignatius’s report in the Washington Post:

Will the Pentagon, with its 30-year planning cycle for building ships, still be launching aircraft carriers in 2048 — even though they’re highly vulnerable to attack today?

That’s an example of the military-modernization questions that kept nagging participants at last weekend’s gathering of the Aspen Strategy Group, which annually brings together top-level current and former national security officials, along with a few journalists, to discuss defense and foreign policy. This year’s focus was on “Maintaining America’s Edge” in the dawning era of high-tech combat, and the big takeaway was this: The future of warfare is now, and China is poised to dominate it.

Speakers at the conference described a new generation of combat systems, powered by artificial intelligence, cyberweapons and robots that can operate on land, sea and air. But America is still largely wedded to legacy weapons of the past — superbly engineered (but super-expensive) aircraft carriers, bombers, fighter jets and submarines.

“We have a small number of exquisite, expensive, manned, hard-to-replace systems that would have been familiar to Dwight D. Eisenhower. They are being overtaken by advanced technology,” argued Christian Brose, staff director of the Senate Armed Services Committee. Instead, he said, the Pentagon needs a large number of inexpensive, unmanned, expendable, autonomous systems that can survive in the new electronic battlespace and overwhelm any potential adversary.

“It is not that we lack money. It is that we are playing a losing game,” Brose contended in a paper presented to the group. “Our competitors are now using advanced technologies to erode our military edge. This situation is becoming increasingly dire.”

Future needs are being drowned out by past practices, because of what Brose’s boss, Armed Services Committee Chairman John McCain (R-Ariz.), has called the “military-industrial-congressional complex.” Brose calculates that in the Pentagon’s initial request for $74 billion in new defense spending in fiscal 2019, only 0.006 percent was targeted for science and technology. The National Science Foundation estimates that in fiscal 2015, only 18 percent of the Pentagon’s research and development budget went to basic, applied and advanced research. Major systems claimed 81 percent.

Even when the Pentagon tries to push innovation, it often stumbles. When Ashton B. Carter was defense secretary under President Barack Obama, he created the Defense Innovation Unit Experimental, or DIUx, with offices in Silicon Valley, Boston and Austin. That operation thrived initially, negotiating 60 defense contracts with start-ups. The program has slowed under the Trump administration, despite support from Defense Secretary Jim Mattis, because it lacks funds and bureaucratic support, warned Christopher Kirchhoff, a former DIUx partner. If Mattis can appoint a strong new DIUx leader soon, maybe he can revive this innovation hub.

The biggest technological challenge discussed here was artificial intelligence. In a few years, these systems have taught themselves to play complex games such as chess and Go better than humans, and to recognize voices and objects better, too. And soon, they will be driving the weapons of combat.

China appears determined to seize this AI “high ground” of future conflict. For the past two years, Chinese companies have won an AI competition for detecting objects. The Chinese are happy for the United States to keep building carriers and bombers, so long as they deploy the more advanced technologies that can disable these systems.

Richard Danzig, a former Navy secretary, published a brilliant paper discussed at the conference warning that as AI systems dominate warfare, they will “introduce important new risks of loss of control.” Humans must be “maximally thoughtful and creative” during design (and plan for failure), because these AI-driven weapons will have accidents and unintended consequences. Wise policymakers must avoid a “Dr. Strangelove” world of unsafe killer robots and doomsday machines.

America’s vulnerability to information warfare was a special topic of concern. One participant recalled a conversation several years ago with a Russian general who taunted him: “You have a cybercommand but no information operations. Don’t you know that information operations are how you take countries down?” . . .

Continue reading.

UPDATE: From Rob May’s InsideAI newsletter:

— Commentary —

This week’s commentary is brought to you by Evanna Hu, CEO of Omelas. (Disclosure, I’m an investor)  After I wrote last week about China’s AI policy and the frameworks for evaluating it, Evanna responded with her point of view.  As an expert in both AI and international affairs, she has a good perspective so I asked her if she would be willing to share.

At the Future of War Conference in Washington, DC this past April, Undersecretary of Defense of Research and Engineering Michael Griffin bluntly acknowledged that China is already winning the AI war. The databacks up the claim, with China filing 8,000 patents relating to AI while the US filed less than 1,000 AI-relevant patents in the same time period. In 2013, the two countries were comparable. Furthermore, unlike adversarial actors, such as China and Russia, and allies, including France, the UK, and the UAE, the US still does not have a comprehensive national AI strategy. Though the Pentagon has established a Joint AI Center and has allocated more money towards the adoption of AI in the National Defense Authorization Act for FY2019, the Center and DoD AI strategy are still in early stages.

Simultaneously, the Committee on Foreign Investment in the United States (CFIUS), a government taskforce that earlier this year blocked the entrance of two Chinese conglomerates, Huawai and ZTE, into the US market, has newly-approved expanded powers. Under the new measures passed by Congress a week ago, the government will regulate funding from foreign origins- but specifically targeting China- in US companies, ranging from corporations all the way down to seed-stage tech startups. This means that theoretically CFIUS can stop startup X who is working on cutting-edge AI technology from receiving funding from Tencent, Alibaba, or any of the $2.4 billion Chinese firms poured into Silicon Valley from January to May of this year. On the flip side, if a company does decide to receive foreign funding above a certain percentage, they will not be able to receive grants or contracts from the US government, including DARPA, SBIR, In-Q-Tel, etc. These measures not only narrow the scope of funding but it also radically reduces addressable markets for US’s emerging technology markets.

If we see the new CFIU measures as the “stick” in the “carrot-and-stick” methodology, it is critical that the “carrot” be developed to help domestic AI companies thrive. While there are already conversations around the best approach at the Pentagon and the White House, the conversation is neither synchronized nor entrepreneur/company-centric. Domestic AI companies want two things: a) increase revenue from sales; and b) access to funding. Whatever form the carrot will be, it needs to address these two main concerns. It is the only way that the US can regain its competitive edge in AI and maintain its number one position in emerging technology.

Evanna Hu is the CEO of Omelas, which uses ML/AI to quantify and assess the online security threat environment. She is also an International Security Fellow at New America, a Washington, DC think tank.

Written by Leisureguy

12 August 2018 at 9:11 am

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.