Page 32 - AI Standards for Global Impact: From Governance to Action
P. 32

AI Standards for Global Impact: From Governance to Action



                  5�4  Computing and AI: Endless frontier and exploration


                  Wang Jian, Director at Zhejiang Lab gave a keynote talk highlighting the following areas where
                  computing and AI will play a key transformative role: 

                  •    Relationship between computing and AI: 
                       o  Computing and AI are deeply interconnected, akin to two sides of the same coin. Alan
                          Turing's early work emphasized computing as a tool to extend human intelligence,
                          much like pen and paper. Today, AI is amplifying human creativity and problem-solving
                          abilities. 
                       o  Computing is not just a subset of computer science but a fundamental discipline
                          parallel to fields like physics and life sciences, forming the foundation upon which AI
                          is built. 
                  •    AI as a transformative tool for science: 

                       o  AI is revolutionizing scientific research by enabling scientists to achieve breakthroughs
                          that would have been impossible otherwise. Examples include projects like GeoGPT,
                          an AI system designed to assist geoscientists in data sharing, research collaboration,
                          and discovery. 
                       o  The Deep Time Digital Earth Initiative and related projects demonstrate how AI can
                          provide innovative tools for researchers, with applications in geoscience, classification,
                          and open science. 

                  •    AI and space exploration: 

                       o  The speaker envisions a future where AI and computing are integral to space
                          exploration. Initiatives like the Three-Body Computing Constellation aim to process
                          multi-source data directly in space, reducing the need to send data back to Earth. 
                       o  Proposed computing satellites would complement existing satellite types
                          (communication, navigation, and observation) to perform advanced in-orbit processing,
                          enabling real-time collaboration for phenomena like gamma-ray bursts and solar
                          observations. 
                  •    AI as a fundamental scientific tool 

                       o  AI is becoming as essential to science as mathematics, applicable across various
                          disciplines. The development of domain-specific foundation models tailored for
                          scientific data (beyond text) will further enhance AI’s role in advancing research and
                          innovation. 
                       o  Shared infrastructures, such as the open platform 02x.org, aim to democratize access
                          to AI capabilities and foster global collaboration among scientists. 
                  •    Governance and safety in AI 

                       o  Governance is critical for ensuring AI and computing technologies are used responsibly.
                          The GeoGPT governance committee is a notable example of how structured oversight
                          can address safety, ethical, and intellectual property (IP) concerns in scientific
                          applications of AI. 
                       o  Open science practices, such as those demonstrated in GeoGPT projects, highlight
                          the importance of transparency and collaboration in advancing technology for the
                          benefit of humanity. 

                  •    Global challenges and collaboration: 

                       o  AI and computing can help address pressing global challenges, such as climate change
                          and disasters stemming from natural hazards, by enabling deeper understanding





                                                           20
   27   28   29   30   31   32   33   34   35   36   37