Artificial intelligence (AI) is changing many industries as intelligent systems assist humans with algorithmic decision making, machine learning, computer vision, natural language processing, chatbots and robotics. Deep learning has been a key development that has helped propel the widespread interest in AI. Given AI is becoming more prevalent and maybe regarded as the new ‘electricity’, it is imperative that human needs, interests and values are centered around all future AI developments. Given Human Computer Interaction (HCI) researchers created user centered design principles and practices based on human factors psychology, they are best placed to work with AI engineers to establish Human-Centered AI (HAI) or Human-AI interaction solutions. This paper presents the challenges that would provide HAI solutions, including the need for explainable AI (XAI) in algorithmic decision making and the need to mitigate automation bias which is when humans over trust and accept computer-based advice. Other ethical HAI challenges are outlined including algorithmic bias and ethical chatbots as well as a discussion on the likely democratisation of AI with the aid of end-user machine learning. The challenges that are discussed herein need addressed in order to properly calibrate human-machine trust and to provide a responsible basis for the future of human-machine symbiosis.
|Number of pages||8|
|Publication status||Accepted/In press - 22 Aug 2019|
|Event||RoCHI 2019: International Conference on Human-Computer Interaction - Romania, Bucharest, Romania|
Duration: 17 Oct 2019 → 18 Oct 2019
Conference number: 16th
|Conference||RoCHI 2019: International Conference on Human-Computer Interaction|
|Period||17/10/19 → 18/10/19|
- human centered AI
- Decision support
- artificial intelligence
Bond, RR., Mulvenna, M., Finlay, D., Wong, A., Koene, A., Brisk, R., Boger, J., & Adel, T. (Accepted/In press). Human Centered Artificial Intelligence: Weaving UX into Algorithmic Decision Making. 2-9. Paper presented at RoCHI 2019: International Conference on Human-Computer Interaction, Bucharest, Romania.