As AI systems have become more powerful and pervasive in society, concerns about AI's potentially negative impacts have dramatically increased. In life-critical situations such as healthcare and security, these issues are particularly important and controversial. At Kitware we have been involved in the research and development of ethical AI frameworks and systems for critical applications, sponsored by the US Department of Defense (DoD) over the past six years. While many would not associate the DoD with ethical AI, in 2021 the DoD published its five Principles of Ethical AI and has expended considerable resources in operationalizing those principles in AI programs across the department. Under DARPA funding across multiple programs, Kitware has developed capabilities that add ethical reasoning to AI systems for battlefield surveillance, medical triage, target detection and others, adhering to the DoD's five EAI principles, the laws of war and the rules of engagement. Methods for explainability, ethical reasoning in LLMs and others will be presented, with quantified results in battlefield surveillance and medical triage domains.
Dr. Anthony Hoogs is the Vice President for AI at Kitware, where he leads Kitware’s Computer Vision Team with more than 60 members including 25 PhD’s. For more than three decades, he has supervised and performed research in various areas of computer vision including: ethical and explainable AI; remote sensing; media forensics; event, activity and behavior recognition; deep learning; object detection, recognition and tracking; and content-based retrieval. He has led dozens of projects, sponsored by commercial companies and government entities including DARPA, AFRL, NGA, ONR, and I-ARPA, that range from basic, academic research to developing advanced prototypes and operation systems. He has been the overall Principal Investigator on multiple large DARPA programs, where he was responsible for overseeing collaborations with more than 25 universities and more than ten commercial subcontractors. Previously at GE Global Research (1998-2007), Dr. Hoogs led a team of researchers in video and imagery analysis on projects sponsored by the US Government, Lockheed Martin and NBC Universal.
Dr. Hoogs received a Ph.D. in Computer and Information Science from the University of Pennsylvania in 1998; an M.S. from the University of Illinois at Urbana-Champaign in 1991; and a B.A. magna cum laude from Amherst College in 1989. He has published more than 100 papers in computer vision, pattern recognition, artificial intelligence and remote sensing. His academic service includes General Chair and Program Chair roles for major computer vision conferences such as the IEEE/CVF Conference on Computer Vision and Pattern Recognition and the IEEE Winter conference on Applications of Computer Vision. He regularly serves as an Area Chair and reviewer for premier computer vision and AI conferences and workshops.