Critical technologies

As economic security is increasingly conflated with national security, the US government must clearly define its strategic objectives to avoid the appearance of protectionism and ensure that controls on advanced technologies are surgical and narrowly targeted to avoid depriving firms of export-based revenue. (p. 12)
Determining criticality depends on dividing technology into specific categories based on specific applications; hardware, software, and other intangible items like knowledge transfer (deemed exports) can all be critical to national security interests, with enforcement of controls of deemed exports representing an inherent challenge. (pp. 12-13)
Quantum technologies can be divided into three subcategories: sensing, computing, and communications; of these, sensing is currently poised for mission use while computing and communications are still emerging. (pp. 13-14)
Quantum technologies can be divided into three subcategories: sensing, computing, and communications; of these, sensing is currently poised for mission use while computing and communications are still emerging. (pp. 13-14)
There is a debate over how to implement export controls of quantum technologies; the industry focuses on qubits, though other thresholds may have value; physical components and proprietary software may also be good bases for controls. (p. 16)
Challenges to placing export controls on quantum technologies include:
1. the need to focus controls narrowly to avoid putting US industries at a competitive disadvantage and restricting access to the large ecosystem needed to develop technology and explore commercial opportunities;
2. determining what metrics to use when identifying thresholds for controls;
3. the multiple ways that quantum computers can be built and that different methods require their own inputs; and
4. the absence of horizontal integration which makes production more expensive now and export controls even more costly. (pp. 16-18)
Instead of controls, the US government should focus on increasing domestic quantum technology capacity. (p. 18)
Semiconductors are both one of the most critical national security technologies and a key part of the civilian economy with highly complex global value chains with many potential chokepoints requiring high levels of international coordination to enforce export controls effectively; the US and Taiwan represent the world’s strongest chokepoints for the most advanced chips. (pp. 18-19)
Chinese military AI systems use US chips, allowing US export controls to degrade Chinese AI capabilities; the US adopted unilateral export controls on advanced AI chips to China, later coordinating with Japanese and Dutch governments; the new controls end the licensing policy based on identifying reliable end-users; the CHIPS and Science Act also sets our guardrails restricting funds from bolstering enterprises that could pose a national security risk. (pp. 19-21)

Concerns with current and future semiconductor export controls include:
1. the economic health of the semiconductor industry and consequently its ability to withstand the costs of additional controls;
2. China’s potential retaliation by weaponizing trade in mid-tier chips which could further depress revenue and cause firms to reduce R&D funding and undermine innovation;
3. the availability of technological workarounds using less advanced equipment that undermine the effectiveness of controls;
4. the disproportionate economic effects of additional controls because of insufficient supply of memory chips in the US and location of supply chains in China;
5. whether China and other countries will develop their own products, “designing out” US technology and allowing their own companies to capture their domestic market share;
6. whether US companies can find alternative markets to make up for the loss of China as a customer; and
7. porousness of controls enforcement. (pp. 21-23)
Long-term costs of the US controls on semiconductors may be larger than expected and become a drag on the US industry’s ability to compete and inspire other entrants into the market; the best approach is to maintain current controls but treat them as a moving target while making clear that controls apply to all end users and uses in China through a policy of denial of license applications. (p. 23)
Artificial Intelligence is generally considered to be computer systems that simulate human-level cognition; the AI supply chain relies on gathering raw data to train AI models; in the modeling phase of development, export controls would apply to data used to train AI models; AI can be categorized as narrow, general purpose, superintelligence, or generative, with narrative and generative currently operational while general and superintelligence AI do not yet exist; all forms could pose threats to national security. (pp. 23-24)

There are few operational definitions of AI in the context of supply chains; October 7 export controls on AI capabilities have focused on hardware components though there is an urgent need to develop controls for intangible goods like algorithms; given the intricacies of controlling this technology, developing uniquely tailored controls is critical; the US should consider building new export control rules for dual-use data flows that feed AI systems. (pp. 24-25)
Biotechnology is important as a force multiplier throughout the tech ecosystem and an important sector for economic growth; the US has deemed leadership in biotechnology as a national security imperative; the US has taken steps to remediate risks of relying on China for biotechnology and increase US investment; biotechnology has been designated as an emerging technology for export controls purposes; controls face the challenge of being strict enough to protect national security but flexible enough to facilitate economic growth. (pp. 25-26)
Concerns over biotechnology include that it could potentially lead to the proliferation of bio-weaponry, or that applications could give bad actors access to DNA or personal health records; designing effective export controls is a complex endeavor as some applications have both significant benefits for civilians and potential for military applications; development and use of biotechnology are restrained by two chokepoints: reliance on people with know-how to develop innovations and access to data necessary to apply biotechnology innovations. (pp. 26-27)
The US should work within the Australia Group to multilateralize controls on bioweapons; expand controls on data and knowledge transfers; and require exporters of personal sensitive biological data to obtain an export license for data flows with national security implications, allowing exemptions for healthcare data to move freely across borders in cases when the health security of the US population is imperiled by pathogens or other risks. (pp. 27-28)
Intangible goods can include data, digital files, software, spyware, and machine learning models; weaponization of data may characterize conflict in the coming century; export controls of intangible goods will likely need to cover the data feeding these applications. (pp. 28-30)