The American Council for Technology-Industry Advisory Council (ACT-IAC) released the its Ethical Application of AI Framework whitepaper on Oct. 8 to help government leaders ethically determine how much confidence to place in technologies that comprise AI.
“This paper addresses the overall ethical framework through an index that evaluates five core parameters underpinning the impact of AI: Bias, Fairness, Transparency, Responsibility, and Interpretability,” the white paper states. “By addressing and incorporating these five components, the framework can be used to understand the ethical automation of solutions to meet desired mission outcomes.”
According to the whitepaper, the framework allows for a more consistent measure on how qualify and quantify components in creating, operating, and improving AI capabilities. The components can then be scored based on applicable indicators that quantify the quality of characteristics, and resulting implications that quantify impact. The components will then produce an overall score for evaluating the ethics of an AI system, to be monitored and checked over time.
“As organizations become more dependent upon the many technologies which comprise AI, there is a need to determine how much confidence and trust to place in them,” the whitepaper states. “To accomplish this ethically, a means to determine its performance and continually monitor for mission veracity or any adverse impacts is needed.”