Human-AI collaborative generative AI for comfortable care: a doctor–patient–AI feedback loop for pre-operative distress and workforce shortages
- 1Yantaishan Hospital
Abstract
Psychological screening is easily displaced by physiological safety checks when preoperative teams have little time and too few staff. Healthcare systems are already affected by widespread workforce shortages and clinician burnout. We suggest that preoperative anxiety consequently often remains unmanaged, with implications for postoperative recovery and hospital workload. Generic educational materials do not adequately address differences in patients’ anxiety, and “health data poverty” can further exclude vulnerable groups. We propose a conceptual human-machine collaborative (HMC) generative artificial intelligence framework to address these clinical and occupational problems together. In the proposed doctor–patient–AI feedback loop, natural language processing would dynamically assess psychosocial risk and guide patient-specific, multimodal education, including dialect-adapted voice agents, as a source of baseline emotional support. Routine interactions would be automated under clinician supervision. Our hypothesis is that this arrangement may conserve clinicians’ limited emotional bandwidth and allow healthier “cognitive reappraisal” in place of chronic fatigue. Implementation science provides a basis for considering how lightweight AI models could be scaled across regional health networks to bring standardized psychoeducation to under-resourced rural areas. This remains a theoretical pathway. Through sociotechnical governance and “data solidarity” to fairly manage algorithmic liability, the framework is still a conceptual framework, but it is operational: it proposes a path to improve perioperative workflows and health equity in the face of severe human resource constraints.
Related
- AI expert collaborative generative model for predicting pathologic response to neoadjuvant chemoimmunotherapy in NSCLC
- Privacy-preserving collaborative transfer learning with generative adversarial networks (CTL-GAN) for classification of dry eye disease
- Training the public health workforce for generative AI: results from a practical workshop in Italy
- Frequency, causes, and duration of drug shortages in South Korea, 2018–2024: a nationwide descriptive analysis of regulatory reporting data and temporal patterns across the COVID-19 pandemic
- Design and validation of a modular smart headband with embroidered electrodes for comfortable EEG monitoring
BibTeX
@article{sun2026humanai,
title = {{Human-AI collaborative generative AI for comfortable care: a doctor–patient–AI feedback loop for pre-operative distress and workforce shortages}},
author = {Hui Sun and Yan Luan and Bo Song},
journal = {Frontiers in Public Health},
year = {2026},
month = oct,
doi = {10.3389/fpubh.2026.2001218},
url = {https://www.frontiersin.org/articles/10.3389/fpubh.2026.2001218/full},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Human-AI collaborative generative AI for comfortable care: a doctor–patient–AI feedback loop for pre-operative distress and workforce shortages
AU - Hui Sun
AU - Yan Luan
AU - Bo Song
PY - 2026
DA - 2026/10/28
JO - Frontiers in Public Health
DO - 10.3389/fpubh.2026.2001218
UR - https://www.frontiersin.org/articles/10.3389/fpubh.2026.2001218/full
AB - Psychological screening is easily displaced by physiological safety checks when preoperative teams have little time and too few staff. Healthcare systems are already affected by widespread workforce shortages and clinician burnout. We suggest that preoperative anxiety consequently often remains unmanaged, with implications for postoperative recovery and hospital workload. Generic educational materials do not adequately address differences in patients’ anxiety, and “health data poverty” can further exclude vulnerable groups. We propose a conceptual human-machine collaborative (HMC) generative artificial intelligence framework to address these clinical and occupational problems together. In the proposed doctor–patient–AI feedback loop, natural language processing would dynamically assess psychosocial risk and guide patient-specific, multimodal education, including dialect-adapted voice agents, as a source of baseline emotional support. Routine interactions would be automated under clinician supervision. Our hypothesis is that this arrangement may conserve clinicians’ limited emotional bandwidth and allow healthier “cognitive reappraisal” in place of chronic fatigue. Implementation science provides a basis for considering how lightweight AI models could be scaled across regional health networks to bring standardized psychoeducation to under-resourced rural areas. This remains a theoretical pathway. Through sociotechnical governance and “data solidarity” to fairly manage algorithmic liability, the framework is still a conceptual framework, but it is operational: it proposes a path to improve perioperative workflows and health equity in the face of severe human resource constraints.
SN - 2296-2565
ER -