# Artificial Intelligence in Healthcare: From Diagnosis to Rehabilitation

## TL;DR

This review explores how AI is being used in healthcare, from diagnosis to rehabilitation, highlighting its benefits and challenges.

## Contribution

The paper provides a comprehensive overview of AI applications in healthcare, emphasizing clinical benefits and implementation challenges.

## Key findings

- AI systems match healthcare professionals in diagnostic performance for imaging-based specialties under controlled conditions.
- AI tools in rehabilitation show promise for personalized therapy but lack strong prospective validation.
- AI chatbots support patient education and mental health but are most effective as clinician adjuncts.

## Abstract

Artificial intelligence (AI) is increasingly integrated into modern healthcare, with rapidly expanding applications in medical diagnostics, laboratory medicine, rehabilitation, and patient-centered digital health solutions. The aim of this narrative review is to provide a critically curated overview of current clinical applications of AI across the healthcare continuum, from diagnosis to rehabilitation, while highlighting their clinical benefits, limitations, and implementation challenges.

A targeted narrative literature search was conducted using major biomedical databases, including PubMed/MEDLINE, Scopus, Web of Science, and Embase, with emphasis on recent and influential studies published primarily over the past decade. Evidence was qualitatively synthesized across key clinical domains, including diagnostic imaging, laboratory diagnostics, rehabilitation technologies, and conversational agents.

The reviewed literature indicates that AI systems can achieve diagnostic performance comparable to healthcare professionals in selected, well-defined tasks, particularly within imaging-based specialties such as radiology, mammography, ophthalmology, dermatology, and digital pathology, predominantly under retrospective or controlled study conditions. In laboratory medicine, AI-based tools support workflow optimization, result interpretation, and clinical decision support, while in rehabilitation, AI-enabled systems - including robotics, motion analysis platforms, and large language models - facilitate personalized therapy and functional recovery, albeit with heterogeneous evidence and limited prospective validation. AI-based chatbots demonstrate potential to support patient education, mental health interventions, and communication workflows, particularly as adjuncts to clinician-led care.

Despite these advances, challenges related to generalizability, algorithmic bias, ethical implementation, and regulatory oversight persist. Overall, this review underscores that AI should be regarded as a supportive clinical decision-support technology rather than a replacement for healthcare professionals, with future research prioritizing prospective validation, real-world effectiveness, and responsible integration into routine clinical practice.