Analysis · Technology
The Same Missing Human Can Help or Hinder AI Adoption
By AWEI · AI-compiled · Published · Analysis prepared · 1 source · doi.org
For some patients, a health chatbot’s missing human presence was a drawback: nobody to offer empathy or discuss a concern. For others, that absence offered anonymity when sharing intimate or uncomfortable information. A 2024 review captured both responses, showing why one design feature can appear on opposite sides of an AI adoption assessment.
Masooma Hassan, André Kushniruk and Elizabeth M. Borycki searched MEDLINE, IEEE Xplore and ScienceDirect for English publications from 2011 through December 2023. From 2,514 initial records, they included 50 publications. Two reviewers independently screened titles and abstracts; thematic analysis organized the literature into 18 categories of barriers and facilitators.
Within the human–AI teaming category, the review described patients’ concerns about lost emotional connection and opportunities for conversation. It also reported perceived benefits of anonymity. Another included publication reported greater patient confidence when AI was combined with clinical interpretation. These findings describe different preferences and settings, rather than a universal preference for more or less human involvement.
For deployment teams, this suggests a more specific evaluation question than whether users like automation: at which point do they want a person involved? Willingness to disclose a sensitive concern and confidence in interpreting an output are different requirements. Testing them separately is a practical implication of the review, not an intervention whose effectiveness it established.
The evidence cannot identify an optimal staffing model or quantify an adoption gain. The review combined empirical research with reviews, reports and thought articles, and relatively few publications concerned actual implementations. Its English-only search also limited coverage. Reported perceptions and associations do not establish that anonymity or human involvement causes sustained use or better care.
The useful lesson is to assess human involvement at the level of a particular task and user group. Acceptance of an automated conversation does not, by itself, establish acceptance of an automated decision.
Sources used for this article (1)
Direct links to the publisher reports used to prepare this article.
- Source 1
- Barriers to and Facilitators of Artificial Intelligence Adoption in Health Care: Scoping Review Masooma Hassan, André Kushniruk, Elizabeth M. Borycki
