Adapt and overcome: Perceptions of adaptive autonomous agents for human-AI teaming

作者：

Highlights：

• Mixed methods study that explored adaptive autonomous teammates on human-AI teams.

• Results showed that work cycles should dictate the ideal autonomy levels over time.

• Interviews revealed that AI teammates should adapt dynamically to seem more human.

• Manual adaptation training modes for team training are critical to acceptance.

摘要

•Mixed methods study that explored adaptive autonomous teammates on human-AI teams.•Results showed that work cycles should dictate the ideal autonomy levels over time.•Interviews revealed that AI teammates should adapt dynamically to seem more human.•Manual adaptation training modes for team training are critical to acceptance.

论文关键词：Autonomous agents,Adaptive autonomy,Human-autonomy teaming,Cyber security,Computer security,Incident response

论文评审过程：Received 9 March 2022, Revised 6 August 2022, Accepted 21 August 2022, Available online 26 August 2022, Version of Record 22 September 2022.

论文官网地址：https://doi.org/10.1016/j.chb.2022.107451