CSET Survey
As CSET continues to grow, we want to learn more about our readers and how we can better serve their interests. If you’re interested in helping, you can do so by filling out a short, anonymous survey. Thank you for your support of CSET and policy.ai!
Worth Knowing
China’s New Data and Algorithms Laws — Plus More PRC Tech News: Recent developments in China — including several major pieces of legislation — will likely affect data management, cybersecurity and AI development globally:
- On August 20, the National People’s Congress passed a new data privacy law, the Personal Information Protection Law, which will go into effect on November 1. Experts say that unlike the European General Data Protection Regulation, to which the PIPL has been compared, the measure is meant to act as a “framework law” that sets out broad objectives and principles with the goal of protecting against government or private sector mishandling of data and personal information.
- On August 27, the Cyberspace Administration of China announced draft regulations on algorithmic recommendation systems. Among other things, the proposed regulations — which would apply to private sector companies, but not government entities — would give individuals the ability to see and edit the tags a platform has assigned to them.
- On September 1, China’s new Data Security Law went into effect. Passed by the NPC in June, the law applies to all companies — both domestic and international — with a presence in China. Some experts have raised concerns that one aspect of the law, a provision that requires companies and researchers within China to disclose information about software vulnerabilities to the Chinese government, could pose a serious cybersecurity risk to non-Chinese entities, including the U.S. government.
- SenseTime Group Ltd., one of China’s largest AI companies, filed for an IPO in Hong Kong that could raise more than $2 billion. The company reportedly partnered with British bank HSBC and China International Capital Corp. for the listing. SenseTime was placed on the U.S. Department of Commerce’s Entity List in 2019 for its role in “China’s campaign of repression” in Xinjiang, severely limiting its ability to work with U.S. firms.
- Last week, President Xi announced plans to launch a Beijing Stock Exchange for “Small and Medium-Sized Enterprises” (companies with fewer than 500 employees). Some observers speculated that the announcement was meant to assuage domestic entrepreneurs and foreign investors after Beijing’s crackdown on a handful of major tech companies, including Didi, Alibaba and Tencent. Others raised the possibility that the new exchange could be used to force companies listed outside China to delist and relist in Beijing.
- More: Timeline: China’s tech crackdown 2021
- TSMC, the world’s largest semiconductor manufacturer, reportedly plans to raise prices on its chips by as much as 20 percent. The increase will apply to the company’s less-advanced chips, such as those used in cars, while its most advanced chips will see a price hike of approximately 10 percent, according to a Wall Street Journal report. The Taiwan-based chipmaker — which manufactures chips for many of the world’s biggest companies, among them Apple, AMD, Nvidia and Qualcomm — dominates the contract foundry market, commanding 54% of global market share last year. While its decision will likely mean higher prices for a number of products, including consumer electronics and the chips used in AI systems, observers say the higher prices could help bring the global chip shortage under control by reducing demand and fueling investment in new chipmaking capacity.
- On Tuesday, the CEO of Intel Corporation, Pat Gelsinger, announced plans to build two new chip fabs in Europe, an investment that could total up to $95 billion over the next decade. Since Gelsinger took over as CEO earlier this year, Intel has doubled down on its semiconductor manufacturing business — in March, the company announced the creation of a standalone foundry business, as well as plans to spend $20 billion on new fabs in Arizona (see our coverage here); in July, Gelsinger met with European leaders to discuss building fabs in Europe, though at the time, the proposed investment was pegged at $20 billion (see our coverage here). The locations of the European fabs have not yet been announced.
- Last month, Cerebras Systems, a Silicon Valley-based startup, unveiled its massive new Wafer Scale Engine 2 processor, which it says will be capable of running AI models with 120 trillion parameters. As the latest AI models approach and surpass the trillion parameter mark, the WSE-2 processor appears well equipped to handle the upcoming generation of large language models. Built using TSMC’s 7 nm process, the processor packs in 850,000 cores across 46,225 square millimeters of silicon — for comparison, it is more than 50 times the size of Nvidia’s top-of-the-line GPU, the A100, which contains approximately 7,300 cores. While Cerebras’s processor is unlikely to end up in any consumer hardware, the company is already attracting interest from government and corporate researchers.
- More: Nvidia seeks EU approval for Arm deal, decision due Oct. 13 | The $150 Million Machine Keeping Moore’s Law Alive
Government Updates
Senate Committee Approves FY 2022 Intelligence Authorization Act: In July, the Senate Intelligence Committee unanimously approved its Fiscal Year 2022 Intelligence Authorization Act. The unclassified portions of the bill, published last month, include several provisions related to AI and emerging technology:
- Sec. 336 requires the Director of National Intelligence to report to Congress on trends in technologies of strategic importance to the United States and areas in which competitors are poised to match or surpass the United States.
- Sec. 340 requires the DNI to develop a plan for establishing a modern digital ecosystem for the development, testing, fielding and updating of AI systems.
- Sec. 343 requires the DNI to report to Congress on the potential to strengthen all-source intelligence integration on foreign cyber threats, with a particular focus on cyber supply chain risks.
- Sec. 352 requires the DNI to submit to Congress a plan to increase cooperation with the intelligence agencies of key democratic partners regarding technological competition with China.
- Sec. 601 requires the president to report to Congress annually with a technology strategy to maintain U.S. leadership in critical and emerging technologies relevant to U.S. national security.
House Armed Services Committee Advances NDAA: Last week, the House Armed Services Committee approved the National Defense Authorization Act for FY 2022 by a vote of 57 to two. The bill includes funding increases and a number of sections related to AI and emerging technology:
- $250 million above the president’s request for AI information and communications technology applied research.
- $257.8 million above the president’s request for JAIC operational systems development.
- Sec. 218 introduces a pilot program for using intermediaries, such as state or local government entities, to connect DOD with technology producers and more quickly move research programs from the prototyping stage to implementation.
- Sec. 6003 specifically directs DOD to establish a “national network for microelectronics research and development,” originally authorized by the FY2021 NDAA.
- Title LII addresses a number of recommendations made by the National Security Commission on Artificial Intelligence, which submitted its report to Congress earlier this year. NSCAI recommendations in the bill include provisions directing the Secretary of Defense to: develop a resourcing plan for a digital ecosystem to enable faster development, testing and fielding of AI systems; designate a chief digital recruiting officer within the Office of the Under Secretary of Defense for Personnel and Readiness; and set AI readiness goals for the Department.
Regulators to Investigate Tesla Autopilot After Crashes: The National Highway Traffic Safety Administration has opened an investigation into Tesla’s “Autopilot” system after a number of crashes that resulted in at least 17 injuries and one death. The investigation will focus on 12 incidents involving Teslas that crashed into emergency vehicles while their driver-assist features were engaged. Based on the investigation’s findings, the NHTSA could force Tesla to recall its cars — the investigation covers 765,000 cars in total — and change how its driver-assist systems work. The company has attracted criticism over the promotion of its driver-assist features, with names like “Autopilot” and “Full Self-Driving” that seem to convey a high degree of autonomous capability, though they are far from fully autonomous. The company’s best systems are still only capable of level 2 autonomy (requiring constant supervision from the driver) and Tesla’s engineers have said they don’t expect to reach the level 5 autonomy (the highest level, requiring no input from a driver under any conditions) any time soon. Tesla must turn over crash data from all 12 accidents by October 22.
In Translation
CSET’s translations of significant foreign language documents on AI
CSET’s translations of significant foreign language documents on AI
PRC Think Tank White Paper: White Paper on Trustworthy Artificial Intelligence. This paper co-authored by a PRC state think tank describes the importance and difficulty of improving the “trustworthiness” of AI systems. The authors recommend increased use of methods such as federated learning and differential privacy to strengthen AI systems’ capability to withstand cyberattacks. The white paper’s policy recommendations include drafting more Chinese legislation related to trustworthy AI, developing commercial AI insurance policies, and taking a cautious approach to research on artificial general intelligence (AGI).
If you have a foreign-language document related to security and emerging technologies that you’d like translated into English, CSET may be able to help! Click here for details.
Job Openings and New Funding
CSET is happy to announce an increase in grant funding, bringing our total funds to more than $100 million and enabling us to continue our work at the intersection of emerging tech and security. With that exciting news comes even more — CSET is hiring! Please apply or share the roles below with candidates in your network:
- AI Research Subgrant (AIRS) Program Director: CSET’s AIRS program will promote the exploration of foundational technical topics that relate to the potential national security implications of AI over the long term via research subgrants. The Director of AIRS will manage all technical, programmatic, and financial aspects of the new AIRS program. Apply now!
- Research Fellow – Cyber/AI: CSET’s CyberAI project is currently seeking Research Fellow candidates to focus on machine learning (ML) applications for cybersecurity to assess their potential and identify recommendations for policymakers (background in ML programming or cybersecurity highly desired:). Complete your application by October 1!
- Senior Fellow: CSET’s Senior Fellows provide mentorship and intellectual leadership; shape and lead lines of inquiry and research projects aligned to our research priorities; and facilitate engagements with government, military, academic, and industry leaders. Apply now!
What’s New at CSET
REPORTS
- Indonesia’s AI Promise in Perspective by Kayla Goode and Heeu Millie Kim
- Responsible and Ethical Military AI: Allies and Allied Perspectives by Zoe Stanley-Lockman
- Headline or Trend Line?: Evaluating Chinese-Russian Collaboration in AI by Margarita Konaev, Andrew Imbrie, Ryan Fedasiuk, Emily Weinstein, Katerina Sedova and James Dunham
- Small Data’s Big AI Potential by Husanjot Chahal, Helen Toner and Ilya Rahkovsky
- Center for International Maritime Security: Leviathan Wakes: China’s Growing Fleet of Autonomous Undersea Vehicles by Ryan Fedasiuk
- 9DASHLINE: A Roadmap for India-U.S. Military AI Collaboration by Husanjot Chahal
- Data Snapshot: Concentrations of AI-related Topics in Research: Computer Vision by Autumn Toney
- Data Snapshot: Concentrations of AI-related Topics in Research: Natural Language Processing by Sara Abdulla
- The Buzz on Bank Automation News: Separating AI cybersecurity hype from reality featuring Micah Musser
Foretell has launched a new project that combines expert and crowd judgment. You can read more about the experts’ views, including how they think trends like China’s military aggression, political polarization, and the strength of the tech sector affect the DOD-Silicon Valley relationship. See all 20 forecast questions associated with this project here.
IN THE NEWS
- Breaking Defense: A Breaking Defense article by Brad D. Williams covered the new CSET brief Headline or Trend Line?: Evaluating Chinese-Russian Collaboration in AI by Margarita Konaev, Andrew Imbrie, Ryan Fedasiuk, Emily Weinstein, Katerina Sedova and James Dunham.
- Breaking Defense: Another article by Williams — this one about China’s new data security law — quoted Research Analyst Dakota Cary, who discussed the implications of the law for China’s offensive cyber operations.
- National Journal: Director of Strategy Helen Toner spoke to National Journal’s Brendan Bordelon for an article about China’s crackdown on domestic tech companies.
- Inside Higher Ed: Elizabeth Redden recapped the CSET brief China is Fast Outpacing U.S. STEM PhD Growth by Remco Zwetsloot, Jack Corrigan, Emily Weinstein, Dahlia Peterson, Diana Gehlhaus and Ryan Fedasiuk in a recent Inside Higher Ed piece.
- TechCrunch: Phil Wagner cited the 2020 issue brief “Cool Projects” or “Expanding the Efficiency of the Murderous American War Machine?”: AI Professionals’ Views on Working With the Department of Defense by Catherine Aiken, Rebecca Kagan and Michael Page in an article about Silicon Valley’s sometimes-fraught relationship with the U.S. military.
What We’re Reading
Article: The Semiconductor Heist Of The Century | Arm China Has Gone Completely Rogue, Operating As An Independent Company With Inhouse IP/R&D, Dylan Patel, SemiAnalysis (August 2021)
Article: Too many AI researchers think real-world problems are not relevant, Hannah Kerner, MIT Tech Review (August 2021)
Report: Facial Recognition Technology: Current and Planned Uses by Federal Agencies, U.S. Government Accountability Office (August 2021)
Article: This is the real story of the Afghan biometric databases abandoned to the Taliban, Eileen Guo and Hikmat Noori, MIT Tech Review (August 2021)
Upcoming Events
- September 16: CSET Webinar, Can AI Write Disinformation?, featuring Andrew Lohn, Katerina Sedova and Micah Musser, moderated by Girish Sastry of OpenAI
- September 16: US-Europe Alliance and CSET, Creating Secure Supply Chains for Transatlantic Innovation, featuring Molly Montgomery, Paul Massaro and CSET’s Will Hunt
What else is going on? Suggest stories, documents to translate & upcoming events here.