AI tools in rehabilitation research: a call for full disclosure

Abstract

Advances in artificial intelligence (AI) are rapidly transforming rehabilitation research, from predictive models of gait and recovery to AI-driven robotics and tele-rehabilitation systems. Such work promises more personalized therapy, but it also raises a question about fully disclosing AI tools in research . Despite the hype, a recent study shows that AI involvement in medical research is often underreported. In a systematic review of 65 clinical trials involving AI,only 10 explicitly mentioned that they used AI, and many AI-specific details were missing[1]. These findings suggest that there is a wide transparency gap.  

Some authors simply may not realize that standard reporting guidelines exist, and others may resist publishing code or details of AI usage. Resnik and Hosseini (2025) reported that some publishers have different and sometimes contradictory policies on AI disclosure [2]. So, without clear directives, many investigators share minimal information. However, whatever the reason is, if the methodology isn’t fully reported, it can’t be independently verified or reproduced. Failing to declare AI methods is not a mere editorial oversight; it poses real risks to science and patients. From a scientific standpoint, transparency is essential for reproducibility.

Finally, not disclosing AI use can backfire academically. Many journals and professional societies now recognize that using AI (even for writing or image generation)can raise questions of authorship and ethics. The ICMJE updated its guidelines to address AI explicitly, noting that manuscripts must disclose information on how work conducted with the assistance of artificial intelligence (AI)technology [3]. So, the publishers are moving toward mandatory disclosure: if you used AI in any substantive way, you must tell readers. Rehabilitation research must get ahead of this curve rather than be caught unprepared.

However, merely having guidelines is not enough; they must be adopted. As noted, several rehabilitation journals have not issued AI-specific author instructions. Specialty fields may assume general reporting standards suffice, but complexity demands explicit mention. We believe The Rehabilitation Journal and peer publications will adopt and consider revising their policies and submission checklists to explicitly ask about the use of any AI tools in this study. If so, describe in detail. Without these directives, underreporting is likely to continue. By embracing clear guidelines and ethical norms today, we can use AI’s potential to improve rehabilitation without compromising the integrity or equity of our science.

DECLARATIONS& STATEMENTS

Ethical Statement 

NotApplicable

AI Use Statement

The authors utilized SciSpace and Semantic Scholar for literature support, and Grammarly to enhance language clarity during manuscriptpreparation. All final content was critically reviewed, verified, and approved by the authors

Consent Statement

Not Applicable

Data Availability Statement 

Not Applicable

Acknowledgments 

None to declare.

Funding Sources                                                          

 None to declare.

Conflicts of Interest 

None to declare.  

REFERENCES

1. MartindaleAPL, Llewellyn CD, de Visser RO, Ng B, Ngai V, Kale AU, et al. Concordance ofrandomised controlled trials for artificial intelligence interventions with theCONSORT-AI reporting guidelines. Nature Communications. 2024;15(1):1619. [CrossRef] [PubMed]
2. ResnikDB, Hosseini M. Disclosing artificial intelligence use in scientific researchand publication: When should disclosure be mandatory, optional, or unnecessary?Accountability in Research.1-13. [CrossRef]
3. YooJH. Defining the Boundaries of AI Use in Scientific Writing: A ComparativeReview of Editorial Policies. J Korean Med Sci. 2025 16;40(23):e187. [CrossRef]