Alphabet launches an AI-powered drug discovery lab, DIU rolls out responsible AI guidelines, and AI investment hits record levels

Sign up
Engage with CSET

As CSET continues to grow, we want to learn more about you and how we can best serve your needs. Please let us know by filling out a short, anonymous survey.

If you want to help us develop our next report, join our team by applying to one of our open positions. 

Worth Knowing

Alphabet Launches AI-Powered Drug Discovery Lab: Earlier this month, Google’s parent company, Alphabet, launched a new commercial research lab that will use AI tools to aid in drug discovery. The company, Isomorphic Labs, will be led by Demis Hassabis, the co-founder and CEO of DeepMind — one of the world’s leading AI research companies, acquired by Google in 2014. Last year, DeepMind successfully modeled protein structures as accurately as more expensive and time-consuming methods in what experts called “a breakthrough of the first order.” In a blog post announcing the launch, Hassabis (who will stay on as CEO of DeepMind) said Isomorphic Labs will build on the protein-modeling breakthrough to “reimagine the entire drug discovery process from first principles with an AI-first approach.” While DeepMind’s history of success might inspire confidence in Alphabet’s new project, AI approaches to medicine have generated a great deal of hype before; so far, most have failed to deliver. But success from Isomorphic Labs could prove a windfall for Alphabet: the global drug discovery technologies market is expected to be worth more than $62 billion by 2030.
Third Quarter Results Show AI and Robotics Spending Surge: Investments in AI and automation set records in Q3 2021, with AI startup funding already topping its 2020 numbers through only nine months:
  • According to CB Insights’ State of AI report, AI startups raised a record $17.9 billion worldwide across 841 deals in the third quarter, the fourth record-breaking quarter in a row and an 8 percent increase over Q2’21. Year-to-date funding measured in at a staggering $50 billion. If current trends hold, 2021’s total AI funding will more than double the record $32.2 billion of 2020. The United States set its own record for AI funding, with U.S. startups attracting $10.4 billion across 324 deals during Q3, by far the most of any country.
  • Robotics sales also set records in the third quarter according to the Association for Advancing Automation, which tracks North American sales. Twenty-nine thousand units worth $1.48 billion have been sold so far this year, an increase of 35 percent over the same period last year. Observers pointed to the ongoing worker shortage as a possible explanation for the surging sales — an argument the 43 percent growth in robot sales to the “Food & Consumer Goods” sector would seem to support — but it remains unclear how much of a role shortages have played.
AI startups haven’t been the only firms to see their valuations surge — as The Wall Street Journal reported earlier this week, since Facebook parent company Meta announced plans to up its spending next year (with as much as $9 billion expected to go towards AI investments), chip companies have seen their stock prices soar: shares of Nvidia and AMD have gone up 30 percent and 20 percent, respectively.  
Government Updates

Annual Army Exercise Shows AI Progress: The second edition of the Army’s Project Convergence exercise —  a six-week “capstone experiment” meant to test the Army’s newest technologies — concluded earlier this month. Unlike the “Scarlet Dragon” exercise we covered last month, which aimed to test how current capabilities could be used in a present-day conflict, Project Convergence is focused on the longer term, testing technologies that the Army hopes to integrate over the coming decade. This year’s event reportedly  involved more than 5,000 personnel (including from the Marines, Navy and Space Force) and tested more than 110 technologies across multiple scenarios. According to Defense One, those technologies included AI-enabled reconnaissance, data sharing between services, mesh networking, augmented-reality IVAS headsets and autonomous tools. Next year’s edition of Project Convergence will likely be even bigger, Defense One reports, expanding to new locations and incorporating new technologies as the military furthers its implementation of its Joint All-Domain Command and Control (JADC2) strategy.

NDAA and USICA in the Legislative Spotlight: With the $1 trillion infrastructure bill passed and signed, Congress is now turning its attention to the annual National Defense Authorization Act. While the House passed its version of the bill in September, the Senate has been slower to move on its version of the NDAA. Earlier this week, Senate Majority Leader Schumer took the first procedural steps to tee up the bill for consideration. In a letter to colleagues last weekend, Sen. Schumer indicated plans to attach the Senate-passed text of the U.S. Innovation and Competition Act — a more than $200 billion bill that aims to bolster U.S. competitiveness with China — to the NDAA. While the Senate passed USICA in June with bipartisan backing, the House pursued its own competing legislation, leaving negotiations at an impasse. After House and Senate members from both parties objected to Sen. Schumer’s plan, potentially jeopardizing progress on the NDAA, he and Speaker Pelosi announced an agreement last night to “immediately begin a bipartisan process” to negotiate a final version of USICA separate from the NDAA that both chambers could pass.

Google Re-Enters Race for DOD Cloud Computing Contract: Google executives said the company is interested in bidding on the Pentagon’s Joint Warfighter Cloud Capability (JWCC) cloud computing contract and will submit a bid if invited. The news comes as somewhat of a surprise — in 2018, Google decided to not renew its contract for the DOD’s Project Maven after employee protests. During that step back from military contracting, the Pentagon awarded Microsoft the highly coveted Joint Enterprise Defense Infrastructure (JEDI) Cloud contract, which would have been worth $10 billion over ten years. But after Amazon and Oracle filed lawsuits alleging undue political meddling behind the selection process, the Pentagon scrapped the JEDI contract altogether, replacing it with the JWCC (see our coverage of that decision from earlier this year). At the time, the DOD singled out Amazon and Microsoft as the top contenders for the new contract, but the head of Google Cloud reportedly met with senior Pentagon officials earlier this month to throw the tech giant’s hat in the ring. The DOD says it plans to award the new contract by April 2022.

DIU Issues Responsible AI Guidelines: The Defense Innovation Unit — the DOD office meant to accelerate the military’s acquisition and adoption of emerging technologies — released new “Responsible AI Guidelines” earlier this week. The guidelines offer a step-by-step process for those involved in DIU acquisitions — commercial vendors, program managers and government partners — to follow when developing AI systems. Last year, the DOD introduced five “Ethical Principles for Artificial Intelligence,” which hold that the military’s AI systems should be responsible, equitable, traceable, reliable, and governable. Earlier this year, Deputy Secretary of Defense Kathleen Hicks reaffirmed the Pentagon’s commitment to those principles and laid out a plan for their implementation (see our coverage of that news here). A DIU report accompanying the guidelines cited Hicks’s memo as an inspiration for the new guidelines. While the RAI guidelines only apply to DIU acquisition programs, a DIU official told Nextgov that the organization hopes its guidelines will be adopted by other agencies and worked into future AI procurement efforts.

In Translation
CSET’s translations of significant foreign language documents on AI

PRC Compute Assessment: White Paper on China’s Computing Power Development Index. This white paper, issued by a PRC state-run think tank, calculates China’s total computing power as 135 Exaflops. The white paper also compares China’s compute with other countries, and compares the computing power of different Chinese provinces and regions.

PRC Standards Development Pronouncement: The Chinese Communist Party Central Committee and the State Council Publish the “National Standardization Development Outline.” This official pronouncement lays out, in broad strokes, China’s approach to the development of standards across a broad variety of fields. The document sets near-term (2025) and medium-term (2035) goals for standardization—both international and domestic—in China, and links standards-setting with virtually every aspect of Chinese society.

PRC AI Standards: Guidelines for the Construction of a National New Generation Artificial Intelligence Standards System. This document, issued by the PRC government’s standards-setting body in 2020, calls for the formulation of specific standards across the full spectrum of basic and applied AI technologies. The standards are particularly detailed in the fields of natural language processing and speech-related applications of AI.

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

Please apply or share the roles below with candidates in your network:
  • Survey Research Analyst: Lead CSET survey design and execution in collaboration with analysts. Please apply by December 1.
  • 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.
  • Research Fellow – Cyber/AI: Conduct research on machine learning (ML) applications for cybersecurity to assess their potential and identify recommendations for policymakers (background in ML programming or cybersecurity highly desired). 
Please visit CSET’s careers page to stay up to date on all active job postings.

What’s New at CSET

REPORTSPUBLICATIONS AND PODCASTSFORECASTS
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.

EVENT RECAPSIN THE NEWS
What We’re Reading

Report: 2021 Report to Congress, U.S.-China Economic and Security Review Commission (November 2021)

Report: The Future of the Digital Order, Jeff Cirillo, Lisa Curtis, Joshua Fitt, Kara Frederick, ​Coby Goldberg, Ilan Goldenberg, Andrea Kendall-Taylor, Megan Lamberth, Martijn Rasser and Dania Torres, CNAS (November 2021)

Article: China has a plan to leapfrog foreign chip makers, Dave Yin, Protocol (November 2021) 

Upcoming Events


What else is going on? Suggest stories, documents to translate & upcoming events here.
Schedule Update

policy.ai will be off the week after Thanksgiving, but don’t worry — we’ll be back on December 9 with all the latest in AI, emerging tech and security policy!