【英文】2021年美国新一届政府创新技术议程（47页）

critical factor in the United States’ economic and  military success has been the achievement of global  leadership in advanced technology; however, the  next administration will inherit the country’s most  tenuous global position in this area since the Second  World War. In today’s Fourth Industrial Revolution,  technological change over the next 30 years will make  the last 30 years look insignificant. The next administration will also deal with a dramatically shifting  global landscape influenced by the long-term effects  of the coronavirus pandemic and a Chinese government that is trying to rapidly erode U.S. technological  advantages through legal and illegal means. Winning  this generation-defining struggle for global leadership  in advanced technology will not just affect the U.S.  economy but will also shape the rest of the century for  the entire world. The next administration must have a  comprehensive technology agenda to spur innovation  in the United States, leverage innovative technologies  within government to better serve citizens, mitigate  the challenges posed by technological disruption, and  work with allies to ensure our democratic values drive  development of these new tools.  

Though artificial intelligence (AI) is just one of  many critical emerging technologies, the blueprint for  achieving global leadership in AI can be a useful guide  for how the next administration could foster innovation across a number of technologies. The explosion of  data and computational capability has made advances  in AI possible; but these resources are concurrently  chokepoints preventing the maturity of the industry.  Continued AI innovation will require large amounts  of data and if the federal government provided more  high-quality data sets to the public, entrepreneurs and  researchers could compete more closely on the quality  of their ideas, rather than their access to proprietary  data sets. Open data does not just advance innovation,  it can also promote equity by reducing one source of  bias in AI—inferior training data. While vetted government data sets will not eliminate bias, this coupled  with investment in digital infrastructure can go a long  way in addressing digital equity. Whether it is increasing access to supercomputing resources for academic  researchers to advance basic knowledge or providing  broadband access so underserved communities can  participate in the digital economy, the United States  will not reach its full AI potential if bright minds are  left behind.