Deep Dive: Chinese perspectives on military AI

What’s new: A CSET report provides an assessment of “Chinese Perspectives on AI and Future Military Capabilities.”

Why this matters: A lot of attention is rightly placed on what senior Chinese military leaders say about artificial intelligence, but this focus is sometimes dismissed as “cheap talk” and therefore not taken seriously. CSET, instead, has reviewed 58 journals written by Chinese military officers, defense engineers, and academics in an effort to more concretely assess Beijing’s AI intentions. 

Key findings extracted from the report:

• Chinese experts believe AI will improve detection, targeting, and strikes against military targets—primarily through intelligent munitions, UAVs, and intelligence, surveillance, and reconnaissance (ISR) software. At the same time, Chinese experts were concerned that a reliance on AI would leave systems vulnerable to cyber and other attacks.

• Chinese experts believe AI will undermine strategic stability, both in general and in terms of China’s deterrence relationship with the US. Chinese experts are particularly concerned that American advances in AI could overwhelm Chinese air defenses, increase the vulnerability of Chinese command and control systems, or reduce China’s time to respond to an imminent attack.

• Chinese experts tend to overestimate US military AI capabilities. Chinese discussions of US AI projects resemble Cold War concerns over a “missile gap” between the United States and the Soviet Union.

• Chinese machine learning engineers face significant hurdles in developing and deploying AI applications, including constraints on the technical literacy of service members and the availability of data and computing power. At the same time, they worry that Chinese advancements in AI could proliferate outside the country, and have taken steps to secure information and prevent technology transfer to the United States.

The report also makes the following policy recommendations:

• Maintain the US AI advantage by constraining the PLA’s access to data, personnel, and critical advanced semiconductors.  This study indicates the PLA faces three major obstacles in adopting AI: limited data access, insufficient technical literacy among personnel, and a dearth of high-end microprocessors. The US, the report argues, has several tools at its disposal to improve US advantages in data, personnel, and hardware, while restricting the Chinese military’s access to the same.

• Avoid arms racing. Rather than immediately pursue the most promising AI capabilities, US military leaders should carefully consider the costs and benefits of developing certain AI systems—for example, an ISR system designed to autonomously track Chinese road-mobile transporter-erector-launchers. It is unclear whether or to what extent China may hold back AI development in response to US. restraint. However, Chinese experts frequently point to their capability gap with the United States as justification for accelerating research and investment in AI and intelligent weaponry.

• Mitigate escalation risk. Maintaining a stable nuclear future will demand that the United States undertake both confidence-building measures and assertive negotiations with Chinese counterparts. US negotiations with China to proscribe certain autonomous weapons or AI capabilities should remain separate from broader efforts to constrain China’s nuclear arsenal.

What we’re thinking: We have to start by acknowledging that this excellent paper is based on open source (meaning, unclassified) information. It is certain that both the US and China are developing what has been called “Assassin’s Mace” — or “black capabilities” — AI tools and weapons that are held in reserve until war begins, to surprise the enemy. That said, this methodology of reviewing key articles and other writings is helpful for cutting through the bluster of official statements and speeches and getting further into what Chinese military planners and experts are really thinking. The whole report is worth your time.

Bipartisan plan for AI and NATSEC being floated

What’s new: As if on cue from our first story, US Representatives Will Hurd (R-Texas) and Robin Kelly (D-Illinois) are teeing up a plan they say will cut off Chinese access to AI-tailored semiconductors and push federal agencies to better deploy AI capabilities.

Why this matters: In a world where Congress can feel more divided than ever, there is a growing consensus on confronting China’s technological ascendence — especially regarding national security relevant advancements. 

Key points:

• Reps. Hurd and Kelly worked with CSET and The Bipartisan Policy Center to produce their report on “Artificial Intelligence and National Security.” 

• The report’s bottom line is that, “Processes to develop and deploy defense and intelligence applications of AI systems must focus on human-machine teaming, trustworthiness, and implementing DOD’s ethical principles for AI.

• “American leadership and advanced technology has been critical to our success since World War II,” says Hurd. “It’s time for Congress to play its role.”

• “I think people in Congress now understand that we need to do more than we have been doing,” says Kelly.

• Hurd and Kelly are now drafting a congressional resolution that captures their ideas and will then move to several specific pieces of legislation. 

What we’re thinking: A number of efforts like this, including the final report from the US National Security Commission on AI, will be finalized over the next several months and 2021 is likely to be a year full of AI-related legislative proposals. 

Two hacks you should know about

What’s new: Hacks targeting a genealogy website and a GPS company have information security professionals worried.

Why this matters: Other high-profile attacks are very likely to follow.

The first attack targeted the genealogy website, GEDmatch.

• Hackers broke into the website and created fake DNA relationships between the site’s users with artificial profiles and with suspected criminals, including murderers and rapists.

• GEDmatch is known for allowing law enforcement access to its records and for playing a helpful role in cracking the Golden State Killer case.

• The cyberattack also caused all 1.4 million GEDmatch users to be opted-in to providing their genetic data for law enforcement use.

• You can read more about the attack here.

The second attack targeted the popular GPS and navigation company Garmin.

• A little over a week ago, Garmin was hit by a ransomware attack — where bad guys encrypt sensitive files and extort money to have them decrypted. 

• As of early this week, the company was still recovering, with portions of its Garmin Connect platform still experiencing sync issues. 

• Experts in this Wired article say the Garmin attack illustrates the gap between big business defenses and ransomware sophistication is narrowing as attackers collect larger and larger “ransoms” which they then invest into more elaborate attacks.

• These groups now have huge amounts to invest in their operations in terms of ramping up their sophistication and scale,” said one security researcher.

• You can read more here.

What we’re thinking: Sometimes, the internet stinks. Also, if you want to better understand how attacks like this happen, checkout Heritage’s new report, “Cybersecurity: National Policies and Practices for Understanding Hacks and Reducing Vulnerabilities.“

Beyond TikTok

[Update: Yesterday, the President issued an Executive Order (EO) banning any US citizen or company from doing business with TikTok parent company, ByteDance. A second EO takes similar action against China’s Tencent, which owns WeChat — a messaging, social media, and payment app that is nearly ubiquitous in China and within Chinese diaspora around the world. Both EOs go into effect on September 20th]

What’s new: US Secretary of State Pompeo says the US is scrutinizing Chinese apps beyond just TikTok.

Why this matters: Many of the concerns that have motivated US actions against Huawei and TikTok are also present with a host of other Chinese companies operating in the US — potentially leading to a much broader crackdown on untrusted Chinese technologies.

Key points: 

• “We want to see untrusted Chinese apps removed from U.S. app stores,” Pompeo told reporters this week, adding “apps like TikTok, WeChat and others are significant threats to the personal data of American citizens, not to mention tools for [the Chinese Communist Party] content censorship.”

• SECSTATE also said the US would like to prevent the holding of US companies’ and citizens’ data on Chinese cloud platforms like Alibaba, Baidu, China Mobile, China Telecom, and Tencent.

• Responding to Pompeo’s statement, one observer said, “This is the US essentially saying they’re going to bifurcate the internet … It’s forcing companies to take sides, to choose a ‘blue’ supply chain or a ‘red’ supply chain.”

What we’re thinking: All together now: “Securing nations means securing networks and securing networks means securing supply chains.” Absent a decisive change in Beijing’s posture, this is as inevitable and necessary as it is disruptive. But a word of caution: we must be careful in confronting untrustworthy companies. Our risk assessments should always focus on clear, discernible threats so that we are able to protect both our citizens and the global free market that undergirds their prosperity. This requires great prudence. 

MIT creates “artificial atoms” for Quantum combo

What’s new: Researchers at MIT have designed a new quantum computing architecture that can rapidly share quantum data between processors while simultaneously performing low-error quantum computations, according to SciTechDaily.

Why this matters: Until now, small-scale quantum processors have been unable to conduct calculations while also transferring data, a capability thought to be critical for realizing mature quantum computing.

Key points:

• This advancement was enabled when scientists created superconducting “artificial atoms” that were connected in a tunable configuration to a microwave transmission line.

• “The system realized by the researchers represents a new regime of light-matter interactions,” according to the article.

• “Unlike models that treat atoms as point-like objects smaller than the wavelength of the light they interact with, the superconducting qubits, or artificial atoms, are essentially large electrical circuits. When coupled with the waveguide, they create a structure as large as the wavelength of the microwave light with which they interact.”

What we’re thinking: We’re going to have to take their word for it. But, this sounds pretty cool.