1、Contents TOC o 1-5 h z AcknowledgementsIII HYPERLINK l bookmark20 o Current Document | An Evolving Industry1 HYPERLINK l bookmark26 o Current Document | The United States in Pole Position5Mcity, Luminar, and Aptiv: Institutional and Technical R&D Case Studies8Drivers Ed. for Machines9 HYPERLINK l bo

2、okmark2 o Current Document | Challenges and Recommendations11Tariffs11Export Controls and Investment Restrictions13Positive Foreign Cooperation14Clear Regulatory Guidelines15Consumer Acceptance18 HYPERLINK l bookmark24 o Current Document | The Checkered Flag21About the Authors23Caporal, Reinsch & Mu

3、rray | IVper second. Difficultto manufacture and expensive, additional improvements in technology are needed before deployment at scale. Lidar consistsofa laser, light scanner, photodetector, processor, computing unit, and specialized GPS receiver.Most lidar firms operate in complex supply chains wi

4、th centers in the United States, Europejapan, and South Korea. Much of the final manufacturing and development of these products occurs in California, where many of the startups focused on lidar are centered. Several other companies focused on automotive lidar also exist in Canada, Europe, China, Ja

5、pan, and Israel. Costs for lidar remain high due to the constant investment in innovation and changes to the product.Luminar, an AV sensor and software vehicle company with offices in Palo Alto, Orlando, and Colorado Springs, works with 12 of the top 15 auto companies on AV development. A number of

6、Luminars partners are foreign companies, including Volvo, Toyota via the Toyota Research Institute, and the Autonomous Intelligent Driving (AID) unit of Audi, which is owned by the Volkswagen Group. Supported by foreign and domestic investment, Luminars lidar technology has made its way from the Uni

7、ted States to the streets of Hamburg in five Volkswagen vehicles approved for road testing in the German city with Level 4 automation enableda landmark test in Germany. A similar story has played out with Velodyne, another American lidar company with mobility partners on every continent (except Anta

8、rctica.Aptiv is another player in the CAV space that operates on a global scale. The company integrates lidar and other sensors with software, connectivity technology, data computation, and machine learning to provide new mobility solutions. It employs over 160,000 workers in 44 countries at manufac

9、turing sights and research centers. Headquartered in Dublin, Aptiv operates more facilities in the United States than any other continent. Its U.S. facilities stretch from tech hotspots like Mountain View and Boston to classic automotive sites like Warren, Ohio, and Detroit.Drivers Ed. for MachinesF

10、or the imaging from lidars, cameras, and radars to be useful, AVs must be equipped with software to process the information. Machine learning technology is key to process the flood of information picked up by an AVs various sensors. That information includes signage, road markings, other vehicles, p

11、edestrians, weather conditions, and more. How to make sense of that data and drive accordingly is second nature for humans, who have a lifetime of experience to apply when behind the wheel. Humans can instantly recognize a slew of variables on the road-a stop sign, yellow light, pedestrian about to

12、cross the street, braking vehicle, or animal in the roadand act accordingly. Getting self-driving vehicles to interpret those variables and make decisions as a human driver would一or make even safer decisionsis a monumental challenge. A vehicles sensors are often referred to as its nervous system, wh

13、ile the computational power to make sense of the data picked up by those sensors and make decisions is referred to as the vehicles brain.Machine learning in this area requires both software engineering, hardware know-how, and on-the-ground testing where vehicles move through various scenarios and le

14、arn toCaporal, Reinsch & Murray | 9make decisions. The willingness of some states, such as California and Arizona, to let companies test A Vs on public roads is an enormous attraction for companies foreign and domestic in the AV space. On the technical side, the United States is an AI leader based o

15、n a range of metrics. Daniel Castro, Michael McLaughlin, and Eline Chivot, Who Is Winning the AI Race: China, the EU or the United States? Center for Data Innovation, August 19, 2019, HYPERLINK s:/ /2019/08/who-is-winning-the-ai-race-china-the-eu-or-the-united-states/ s:/ /2019/08/who-is- winning-th

16、e-ai-race-china-the-eu-or-the-united-states/. It is home to over 1,000 companies focused on artificial intelligence, including global leaders like Google (which has its own AV play in Waymo), Facebook, Amazon, Microsoft, and IBM. The United States is flush with venture capital cash and talent in the

17、 AI space as well, is a long-time leader in the hardware that powers machine learning and leads in other areas. OECD, “Private Equity Investment in Artificial Investment/1 OECD Going DIGITAL Policy Note, December 2018, HYPERLINK s:/ /going-digital/ai/private-equity-investment-in-artificial-intellige

18、nce.pdf s:/ /going-digital/ai/private-equity-investment-in-artificial-intelligence.pdf.The United States is dominating the autonomous vehicle AI space. Prior to 2015, AV firms made up less than 5 percent of total investment in AI start-ups. By mid-2018, AV firms accounted for 30 percent of total inv

19、estment in AI start-ups, with 80 percent of that funding going to U.S. companies. Ibid10 | Fueling the Race: Foreign Investment in Advanced Vehicle Technology Toyota is one of many companies that has embraced一 and in turn supportsthe fusion of the U.S. AI ecosystem with the classic automotive excell

20、ence found in the United States. Toyota operates a venture capital arm, Toyota AI Ventures, as a subsidiary of Toyota Research Institute, which is devoted to applying AI to vehicles. Toyota AI Ventures has invested in a range of cutting-edge U.S. firms, most of which are tackling the challenges at t

21、he core of ACE vehicles.Despite the innovation, significant investment, and excitement surrounding ACE vehicles, the race to electrified and connected vehicles with Level 4 automation (highly automated vehicles capable ofperforming all driving tasks under certain conditions, with the option for huma

22、n override) and above is still in an early phase. Governments are only beginning to grapple with how to regulate AVs, businesses with few small exceptions remain squarely focused on R&D and limited deployment of vehicles that exhibit Level 3 conditional automation, and the general public remains war

23、y of AVs. The United States has time to address those challenges. The competitive advantages discussed above have given it a lead over competitors as the premier destination for ACE vehicle R&D. Further, the average vehicle age in the United States is over 11 years, which provides the private sector

24、 and government ample time to prepare for the automated, connected, and electric future of mobility. As in any long-distance race, endurance, stability, situational awareness, and a well-thought-out game plan are keys to success.3 | Challenges and RecommendationsWhile the United States is already an

25、 attractive place for foreign and domestic companies alike to research, develop, test, and even deploy ACE vehicles, a number of speedbumps remain that if left unaddressed or mishandled could risk the United States position as a global leader in ACE development. The following list of challenges is n

26、ot exhaustive, nor are the recommendations the only right answers to the challenges presented. Rather, this section is intended to highlight some of the many significant issues government and businesses should have on their radar as ACE vehicles continue to develop and approach wide-scale commercial

27、 deployment.Generally, the United States should focus on policies that increase opportunities for cooperation and competition among foreign and domestic companies, and it should not adopt policies that restrict trade, investment, and collaboration among industry stakeholders.TariffsTHE ISSUE:In May,

28、 the Trump administration concluded in its Section 232 investigation that U.S. imports of automobiles and auto parts pose a threat to U.S. national security. White House, “Adjusting Imports of Automobiles and Automobile Parts Into the United States,“ Presidential Proclamation, May 17, 2019, HYPERLIN

29、K s:/ /presidential-actions/adjusting-imports-automo-biles-automobile-parts-united-states/ s:/ /presidential-actions/adjusting-imports-automo- biles-automobile-parts-united-states/. Domestic and foreign automakers argue that imports do not pose a risk to U.S. national security and that tariffs or ot

30、her restrictions would do more to harm than help the industry. Section 232 National Security Investigation of Imports of Automobiles and Automotive Parts Hearing Panel,“ July 19, 2018, HYPERLINK s:/ /contentStreamer7documentId=DOC-2018-0002-2299&content%e3%83%bb s:/ /contentStreamer7documentId=DOC-2

31、018-0002-2299&content Type=pdf.Caporal, Reinsch & Murray 111 The automakers strongly oppose trade restrictions on automobiles and parts. The administration found that rising automotive imports alongside protected foreign markets in the European Union and Japan have cut into domestic automakers, reve

32、nue and therefore, their ability to invest in R&D. According to the administration:Sales revenue enables R&D expenditures that are necessary for long-term automotive technological superiority, and automotive technological superiority is essential for the national defense. The lag in R&D expenditures

33、 by American-owned producers is weakening innovation and, accordingly, threatening to impair our national security.36Under that basis, the Secretary of Commerce concluded that automobile and parts imports threaten U.S. national security; however, the president has not exercised his authority provide

34、d under Section 232 of the Trade Expansion Act of 1962 to restrict imports. Regardless, the administration made clear that it believes investment and the innovation contribution from U.S.-owned automotive companies is more valuable than that of foreign automakers. Further, linking automobile R&D and

35、 production to national security does not inherently mean that restricting imports would be advantageous for U.S. national security or that more capacity is required to meet defense needs. In addition, the overwhelming majority of U.S. auto imports come from allies such as Canada, Mexico, Japan, Sou

36、th Korea, and the European Union. The administrations rhetoric and reasoning alone can have a chilling effect on foreign automakers and companies considering investment in the United States.37Ironically, tariffs would likely damage both domestic and foreign auto companies, choke revenue for R&D, and

37、 encourage the entire industry to shift at least some R&D and production outside of the United States.38 Imports of motor vehicles and parts from the European Union and Japan amounted to almost $120 billion in 2018, which is about one-third of all U.S. auto and parts imports. Vehicles produced in th

38、e United States for domestic consumption or export abroad rely on foreign parts and global value chains to maintain a competitive position in the global market. Foreign retaliation would put further pressure on automakers in the United States. Higher prices of autos assembled in the United States as

39、 a result of tariffs would discourage domestic sales and make U.S. assembled autos less globally competitive. Lower demand due to higher vehicle prices would result in less production, fewer jobs, and less revenue for R&D.Tariffs on components necessary for ACE vehicles, such as batteries, sensors a

40、nd other lidar components, and chips, would deter R&D in the automotive and tech sectors. Tariffs on parts that make up the nervous system and brain of A Vs ignore the reality that those parts rely on softwareusually developed in the United States or at least in partto have any real utility or value

41、. Sensors and chips have limited utility without cutting-edge software to instruct and teach A Vs how to interpret and react to data gathered. Tariffs used in a bid to protect the United States ability to competitively manufacture certain AV parts do not play to the U.S. strengths outlined aboveits

42、high rate of graduates in the information technology space and its lead in software engineering, semiconductor technology, and artificial intelligence. Competitive access to parts integral to ACE vehicles up and down the value chain allows the United States to exploit its own competitive advantages.

43、 Cutting off access would minimize those advantages. Tariffs on ACE components would also complicate the ability of automakers and retailers to keep vehicle prices relatively constant despite customers expectation that new features continue to be added to vehicles. Tariffs could make strategic partn

44、erships between U.S. and foreign automakers and other players in the ACE vehicle space untenable as well.In short, tariffs would put the United States position as a global leader in ACE vehicle innovation at risk.12 | Fueling the Race: Foreign Investment in Advanced Vehicle TechnologyRECOMMENDATIONS

45、:The administration should make dear that it will not impose or threaten tariffs or quotas via Section 232 on automobiles and parts, including those necessary for ACE vehicle R&D.If foreign investment raises national security risks, the administration should use carefully tailored, narrow tools to n

46、egate those risks, including the Committee on Foreign Investment in the United States.Export Controls and Investment RestrictionsTHE ISSUE:The Department of Commerce is in the process of reviewing a range of technologies for export controls some of which are at the heart of ACE vehicles. Separate fr

47、om the export control effort, the administration has linked automotive R&D to national security, writing that the rapid application of commercial breakthroughs in automobile technology is necessary for the United States to retain competitive military advantage and meet new defense requirements/ Ibid

48、., 35.Caporal, Reinsch & Murray 113Unilateral U.S. controls that unintentionally capture civilian applications of emerging technologies could frustrate R&D in the United States and risk offshoring R&D to environments where technology and information sharing across borders is less constrained. Unilat

49、eral controls on technologies that can be researched, developed, deployed, and acquired outside the United States would also do little to slow the uptake of those technologies by adversaries. Broad export controls could also limit the amount and quality of human capital companies in the United State

50、s can attract for R&D. Engineers and software developers are increasingly sought after as the economy increasingly becomes digitized. Export controls that limit or complicate the ability of companies in the United States to attract the best and the brightest will result in top global talent being si

51、phoned away to Europe, Japan, China, and elsewhere.Technologies under review for control by the Department of Commerce include artificial intelligence and machine learning, a broad category that includes areas that ACE vehicles will rely on and where U.S. and foreign companies are cooperating on and

52、 competing against each other, such as computer vision, expert systems, speech and audio processing, natural language processing； cloud technologies, and AI chipsets. Other broad categories being considered for export controls that ACE vehicles will rely on include position, navigation, and timing t

53、echnology; microprocessor technology; advanced computing technology; data analytics technology; and logistics technology.The nature of these technologies makes export controls, particularly unilateral controls, complicated. The technologies themselves are not inherently connected to national securit

54、y; in fact, many of them share roots within academia and open-source research. National security implications arise via the application of the technologymilitary platforms enabled or augmented by AI; military logistics made more efficient with AI; unpiloted military intelligence, surveillance, and r

55、econnaissance systems that make use of autonomous navigation and computer vision; and Ai-assisted data processing to parse foreign capabilities or targets.In addition to dual-use applications, these technologies and their research, development, and deployment in the automotive sector are not “made i

56、n America/ and they are not developed by only U.S. companies. Nor are these technologies worked on only by Americans, even in the United States. Foreign talent, investment, and corporate buy-in to develop these technologies within the United States could be put at risk by onerous export controls tha

57、t limit civilian applications of emerging technologies along with military applications. In the near term, overly broad export controls on emerging and foundation technologies could limit U.S. exports of cutting-edge parts and vehicles, deter international R&D within U.S. companies and among partner

58、ships between U.S. companies and foreign companies, and limit access to human capitalall while boosting the competitiveness of companies abroad.RECOMMENDATIONS:U.S. regulators should identify discrete products, rather than systems technology, to control. Those products should exhibit a specific nati

59、onal security dimension. The administration should continue to emphasize end-use in export control decisionmaking to avoid capturing entire systems or technologies.Consideration of license applications for controlled items should follow the Department of Commerces established end-user-based approach

60、 rather than providing for blanket denials.The administration should work with existing multilateral export control regimes to ensure they are up to date with respect to emerging and foundational technologies and that member nations are developing parallel policies.The administration should continue