Capability-Adaptive Intelligent Assistance in Industry 5.0 Manufacturing: A Systematic Review and Research Framework
This systematic review of 14,346 Industry 5.0 records reveals a critical disconnect between the field's dominant technology-centric focus and the under-synthesized literature on worker capability diversity, prompting the proposal of the Capability-Adaptive Intelligent Assistance Framework (CAIA-5.0) and a six-item research agenda to bridge this gap and enable truly inclusive manufacturing.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
The factories of the future are being designed with a new promise: that technology should serve people, not the other way around. This vision, known as Industry 5.0, moves beyond the cold efficiency of the past to focus on human well-being, environmental care, and the ability to bounce back from disruptions. At the heart of this shift is the idea of intelligent assistance. Imagine a factory floor where machines do not just follow rigid commands but can see, understand, and adapt to the human worker standing beside them. These systems use artificial intelligence to perceive their surroundings, digital twins to create virtual copies of the physical world for testing, and extended reality to overlay helpful information directly onto a worker's view. The goal is a seamless partnership where humans and robots collaborate safely and effectively. However, for this partnership to be truly successful, the technology must work for every single person who steps onto the floor, regardless of their physical abilities, age, or how their brain processes information. If the tools are built only for a standard, average worker, the promise of a human-centric future remains unfulfilled.
A recent systematic review set out to examine whether the scientific literature supporting this vision actually accounts for the vast diversity of the human workforce. The researchers analyzed a massive collection of 14,346 records from a major scientific database, covering work published between 2016 and 2026. Their investigation revealed a striking imbalance in how the field is currently developing. The vast majority of the research focuses heavily on the machinery and the code behind it. More than half of the records discuss robotics and collaborative robots, while over a quarter focus on artificial intelligence and machine learning. In contrast, the human side of the equation receives far less attention. Only about 11 percent of the records address human factors, and just 10 percent discuss how people interact with these systems. The most telling finding concerns the specific needs of workers with different capabilities. While a small but real body of work exists that explicitly considers disabilities, age, neurodiversity, or digital literacy, these topics are almost invisible in the broader conversation. When the researchers looked at how often these concepts appeared together in the keywords of the papers, they found that disability-related terms barely registered, appearing together fewer than three times. This stands in sharp contrast to the most common pair of keywords in the entire field, which appeared together 425 times.
The study concludes that while research on accommodating different human capabilities does exist, it is scattered, rarely cited by others, and disconnected from the main stream of technology development. The literature is not missing the work entirely, but it has not been brought together into a coherent whole. The authors argue that this fragmentation prevents the industry from building truly inclusive systems. To address this, they propose a new structure called the Capability-Adaptive Intelligent Assistance Framework. This framework aims to connect the scattered research on human diversity with the powerful, existing tools of artificial intelligence. It suggests a way for systems to adjust their behavior based on the specific needs of the worker, such as changing how much autonomy a robot has depending on the human's current state. The researchers note that one specific part of this idea, which involves automatically adjusting a robot's independence based on a worker's condition, has not yet been tested in practice. They have outlined a clear path forward, calling for six specific areas of research to turn this fragmented collection of ideas into a practical, working guide for designers and policymakers. The ultimate goal is to ensure that the intelligent factories of tomorrow are built not just for the ideal worker, but for the real, diverse people who will actually use them.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.