Interview
How emerging technologies could affect national security | Helen Toner
- Helen Toner serves as Director of Strategy at the Center for Security and Emerging Technology (CSET) at Georgetown University, a think tank established with a $55 million grant from the Open Philanthropy Project, its largest grant to date.
- CSET's primary mission is to produce high-quality analysis on the intersection of emerging technology and national security, with an immediate focus on AI's impact on military operations, intelligence, and US-China relations.
- Toner testified before the US-China Economic and Security Review Commission, emphasizing that AI is a general-purpose technology akin to "advanced statistics" rather than a discrete hardware item like a rocket, making containment policies ineffective.
- A key argument presented to US policymakers is that attempting to restrict AI research through export controls or immigration bans will likely fail because the technology is open-source and widely distributed, potentially driving top talent to Canada, the UK, or home countries.
- Toner advocates for the US to leverage its liberal democratic values as a strategic asset rather than attempting to mimic China's state-led model, noting that the US benefits from a massive net import of global talent (approx. 190,000 inventors/patent holders annually) compared to China's net export.
- The AI "arms race" is described by Toner as a poor analogy to nuclear weapons; AI's strategic impact is expected to be a gradual infrastructure shift (similar to electrification) affecting logistics, data flow, and cyber capabilities rather than discrete, countable weapons systems.
- Toner predicts AI will first significantly impact national security through cybersecurity and automation of hacking rather than autonomous physical weapons, due to the slower procurement cycles for military hardware versus the rapid evolution of software.
- The analogy of "data is the new oil" is rejected as misleading because data is non-rival and application-specific, unlike oil; a country's population size does not automatically translate to a military AI advantage if the specific data required for defense applications is not accessible.
- CSET is actively hiring research fellows and data scientists to analyze open-source data on AI talent flows, investment, and publication trends to provide grounded policy recommendations to US agencies.
- Specific policy recommendations being developed include reforms to H-1B and Green Card processes to improve talent retention, and the creation of a national AI testbed by NIST to standardize safety and reliability evaluations.
- Toner highlights that the Chinese government's primary strategic goal is domestic stability and population control, which may consume resources that would otherwise be directed toward external expansion or aggressive military adventurism.
- The US-China technology competition is framed not as a binary "AI race" but as a subset of broader geopolitical friction, where US strategy should focus on establishing fair global trade rules rather than zero-sum containment.
- Toner notes a significant "credibility gap" when the US mixes trade protectionism with national security claims (e.g., against Huawei), which undermines the ability to make genuine security alerts about foreign technology.
- Regarding the "China Specialist" career path, Toner advises that true expertise requires immersion in the country, ideally through a year or more of language study and residence, rather than superficial familiarity.
- CSET aims to bridge the gap between technical AI knowledge and national security policy by training the next generation of analysts who understand both machine learning fundamentals and the bureaucratic realities of the US defense establishment.
- Potential career entry points identified include graduate programs in security studies (e.g., Georgetown, SAIS), the AAAS Science & Technology Policy Fellowship, and Tech Congress, alongside gaining practical experience in government agencies.
- Toner expressed skepticism regarding the Western media's exaggeration of the "Social Credit System," arguing it is currently in a prototype phase and is merely one small component of a much larger, more concerning surveillance infrastructure in China.
- Living in Beijing for nine months revealed that the expatriate community is diverse, ranging from highly accomplished researchers to those seeking alternative lifestyles, and that domestic tourism significantly alters the interaction between locals and foreigners compared to international tourism norms.
- Toner maintains that expressing views critical of Chinese policies (e.g., regarding Hong Kong or Xinjiang) is necessary given the severity of human rights issues, despite the potential impact on visa access or security clearance eligibility.
- The US government is urged to shift from a "supremacy" mindset to an "interest-based" approach regarding China, recognizing that accommodation and new international equilibria may be necessary as China's economic and military power grows.
- CSET expects to see AI implementation in government lag behind private sector innovation, with early adoption likely occurring in back-office functions (logistics, HR, finance) before mission-critical defense applications.
- The podcast notes that the job market for operations management in high-impact nonprofits has become more competitive in the past year, with 113 such roles currently listed on the 80,000 Hours job board, alongside 70 roles in AI strategy and governance.