The rise of artificial intelligence (AI) challenges us to explore whether human-to-human relationships can extend to AI, potentially reshaping the future of coaching. The purpose of this study was to examine client perceptions of being coached by a simulated AI coach, who was embodied as a vocally conversational live-motion avatar, compared to client perceptions of a human coach. It explored if and how client ratings of coaching process measures and outcome measures aligned between the two coach treatments. In this mixed methods randomized controlled trial (RCT), 81 graduate students enrolled in the study and identified a personally relevant goal to pursue.
The study deployed an alternative-treatments between-subjects design, with one-third of participants receiving coaching from simulated AI coaches, another third engaging with seasoned human coaches, and the rest forming the control group. Both treatment groups had one 60-minute session guided by the CLEAR (contract, listen, explore, action, review) coaching model to support each person to gain clarity about their goal and identify specific behaviors that could help each make progress towards their goal. Quantitative data were captured through three surveys and qualitative input was captured through open-ended survey questions and 27 debrief interviews.
The study utilized a Wizard of Oz technique from human-computer interaction research, ingeniously designed to sidestep the rapid obsolescence of technology by simulating an advanced AI coaching experience where participants unknowingly interacted with professional human coaches, enabling the assessment of responses to AI coaching in the absence of fully developed autonomous AI systems. The aim was to glean insights into client reactions to a future, fully autonomous AI with the expert capabilities of a human coach.
Contrary to expectations from previous literature, participants did not rate professional human coaches higher than simulated AI coaches in terms of working alliance, session value, or outcomes, which included self-rated competence and goal achievement. In fact, both coached groups made significant progress compared to the control group, with participants convincingly engaging with their respective coaches, as confirmed by a novel believability index.
The findings challenge prevailing assumptions about human uniqueness in relation to technology. The rapid advancement of AI suggests a revolutionary shift in coaching, where AI could take on a central and surprisingly effective role, redefining what we thought only human coaches could do and reshaping their role in the age of AI.
| Identifer | oai:union.ndltd.org:columbia.edu/oai:academiccommons.columbia.edu:10.7916/rszn-m994 |
| Date | January 2024 |
| Creators | Barger, Amber |
| Source Sets | Columbia University |
| Language | English |
| Detected Language | English |
| Type | Theses |
Page generated in 0.0172 seconds