Presentation
Why Healthcare Still Lags in AI: Lessons from Industries That Moved Faster
SessionPoster Session 1
DescriptionThe integration of Artificial Intelligence (AI) in different industries has leveraged the role of Human Factors (HF) in ensuring safe, effective, and trustworthy human-machine collaboration. Domains such as aviation, manufacturing, and driving have integrated AI more rapidly, offering valuable lessons. In contrast, healthcare remains below the global average in technology adoption. For instance, between 2015–2018, US healthcare job positions showed one of the lowest demands for AI skills, with only 1 in 1,250 hospital roles requiring AI expertise, compared to much higher rates in information, finance, and manufacturing. Additionally, global AI investment in healthcare grew by 233% between 2020 and 2023, and it is projected to keep expanding to around $200 billion by 2030. However, many hospitals lag not just in AI hiring and investment but in more basic digital foundational capacity. For example, over half of hospitals operate with broadband speeds below 500 Mbit/s, while critical workflows remain paper-based and clinical data are often fragmented across multiple EMRs. Even among high-maturity systems, many lack fully enabled analytics within their EHRs, meaning the underlying data infrastructure is frequently insufficient to leverage AI meaningfully. Therefore, it is important to account for responsible AI adoption in healthcare, ensuring systems are designed with human factors principles. This foundation will allow a smoother integration of AI applications that have already been explored in healthcare, such as clinical decision support, diagnosis, and patient monitoring.
This work aims to address the following question: What are the main advantages that other industries have gained from incorporating AI? What can healthcare learn from different domains in embedding AI? and what role did human factors play in ensuring AI adoption? To explore this, we conducted a targeted review across three domains: aviation, driving, and manufacturing. We prioritized systematic reviews and highly cited case studies published after 2015 in major databases (Wos/Scopus/IEEE), selecting up to four to five papers per domain to balance breadth with depth. Results were summarized in the following structure: human factors approach, AI applications, advantages, disadvantages, and limitations. The analysis is structured, but not limited, around the following human factors considerations in AI applications: trust, workflow integration, safety, and training. Mapping these dimensions in mature industries provides a comparative framework that brings transferable lessons for healthcare. Therefore, the contribution of this work is to synthesize evidence across domains and translate it into strategies for healthcare in AI adoption.
Preliminary results from the reviewed literature showed some patterns across domains. In aviation, AI has improved safety and efficiency through decision support and automation, but concerns remain about trust, explainability, and certification. New systems are often tested in high-fidelity simulators, which helps to build confidence and prepare operators before implementation. In driving, AI has advanced human–machine interfaces and partially autonomous functions to reduce human error, but challenges include driver over-reliance, poor takeover performance, and mixed user acceptance. Findings on training are inconclusive, as most progress has been made in human-centered interface design. In manufacturing, the literature cautions about a narrow productivity perspective, as AI has been mainly applied to ergonomics and workload management, supporting risk and fatigue detection as well as productivity improvements. However, less has been done on cognitive ergonomics, worker training, and a regulatory framework.
Therefore, in healthcare, AI systems will need robust certification pipelines and explainability to support trust and reduce automation-related disruptions. AI must be integrated into clinical workflows to reduce workload rather than add to it, with greater attention on cognitive demands to avoid alert fatigue. Simulation-based training should prepare clinicians for failure modes and handovers. Clear safety and transparency standards will be essential to support evaluation, accountability, and acceptance. At the same time, many healthcare environments are resource-strained, where data foundations are weak and coding practices are inconsistent. When underlying data is incomplete or not meaningful, the risks of downstream errors and unsafe decisions increase substantially, further complicating safe AI integration. Finally, as a limitation, this synthesis is not exhaustive, since it was drawn on a small but targeted set of systematic reviews and case studies across domains.
This work aims to address the following question: What are the main advantages that other industries have gained from incorporating AI? What can healthcare learn from different domains in embedding AI? and what role did human factors play in ensuring AI adoption? To explore this, we conducted a targeted review across three domains: aviation, driving, and manufacturing. We prioritized systematic reviews and highly cited case studies published after 2015 in major databases (Wos/Scopus/IEEE), selecting up to four to five papers per domain to balance breadth with depth. Results were summarized in the following structure: human factors approach, AI applications, advantages, disadvantages, and limitations. The analysis is structured, but not limited, around the following human factors considerations in AI applications: trust, workflow integration, safety, and training. Mapping these dimensions in mature industries provides a comparative framework that brings transferable lessons for healthcare. Therefore, the contribution of this work is to synthesize evidence across domains and translate it into strategies for healthcare in AI adoption.
Preliminary results from the reviewed literature showed some patterns across domains. In aviation, AI has improved safety and efficiency through decision support and automation, but concerns remain about trust, explainability, and certification. New systems are often tested in high-fidelity simulators, which helps to build confidence and prepare operators before implementation. In driving, AI has advanced human–machine interfaces and partially autonomous functions to reduce human error, but challenges include driver over-reliance, poor takeover performance, and mixed user acceptance. Findings on training are inconclusive, as most progress has been made in human-centered interface design. In manufacturing, the literature cautions about a narrow productivity perspective, as AI has been mainly applied to ergonomics and workload management, supporting risk and fatigue detection as well as productivity improvements. However, less has been done on cognitive ergonomics, worker training, and a regulatory framework.
Therefore, in healthcare, AI systems will need robust certification pipelines and explainability to support trust and reduce automation-related disruptions. AI must be integrated into clinical workflows to reduce workload rather than add to it, with greater attention on cognitive demands to avoid alert fatigue. Simulation-based training should prepare clinicians for failure modes and handovers. Clear safety and transparency standards will be essential to support evaluation, accountability, and acceptance. At the same time, many healthcare environments are resource-strained, where data foundations are weak and coding practices are inconsistent. When underlying data is incomplete or not meaningful, the risks of downstream errors and unsafe decisions increase substantially, further complicating safe AI integration. Finally, as a limitation, this synthesis is not exhaustive, since it was drawn on a small but targeted set of systematic reviews and case studies across domains.
Authors
PhD student in industrial Engineering
Event Type
Poster Presentation
TimeMonday, March 234:45pm - 6:15pm EDT
LocationRhinelander Gallery
Digital Health