人类决策与算法决策：重新思考不断变化的范式下的监管和政策方法.pdf

1、Fly AI Forum 2025Human vs.Algorithmic Decision-Making:Rethinking Regulatory and Policy Approaches for a Changing ParadigmSbastien Colmant-23 April 2025F L Y A I 2 0 2 5 F O R U M 6319872025F L Y A I 2 0 2 5 F O R U M 64Setting the Stage:The Evolution of Aviation SecurityWhilst technology did evolve

2、over the 20 years:Human decision-making has been,and still is,the cornerstone of aviation security(avsec).Algorithms have entered the picture but have not fully displaced human judgment.The“gold standard”is still based on human decisions,with algorithms seen as a supplementary toolF L Y A I 2 0 2 5

3、F O R U M 65Setting the Stage:The Evolution of Aviation SecurityU.S.Congress,Office of Technology Assessment,Technology Against Terrorism:Structuring Security,1992.F L Y A I 2 0 2 5 F O R U M 66Repetitive Tasks in AVSEC U S Congress Off?F L Y A I 2 0 2 5 F O R U M?Consequences 68?F L Y A I 2 0 2 5 F

4、 O R U M Consequences 69F L Y A I 2 0 2 5 F O R U M?F L Y A I 2 0 2 5 F O R U M?F L Y A I 2 0 2 5 F O R U M A fast journey to Regulatory adoption!?F L Y A I 2 0 2 5 F O R U M But not yet a reality in pax checkpoint!?Regulators Resistance to AI in Aviation SecurityF L Y A I 2 0 2 5 F O R U M 73?AVSEC

5、 is far away from AIF L Y A I 2 0 2 5 F O R U M 74?Rethinking Regulatory Frameworks for AI in Aviation SecurityNeed for a paradigm shift in regulatory thinking:from human-centric to hybrid AI-human decision-making.Build trust in AI by providing clear,transparent guidelines on AI decision-making proc

6、esses and accountability.Implement AI-based decision-making in phases,starting with lower-risk areas to build confidence.Governments must incentivize AI-driven initiatives and research in aviation security.F L Y A I 2 0 2 5 F O R U M 75Conc