Artificial intelligence, machine learning and deep learning: Potential resources for the infection clinician.

Area of Science:

• Infectious Diseases
• Medical Informatics
• Computational Biology

Background:

• Artificial intelligence (AI), machine learning, and deep learning are increasingly explored for human infection research and management.
• Generative AI is also being investigated for its potential applications in this field.

Purpose of the Study:

• To summarize recent and potential future applications of AI in clinical infection practice.
• To assess the relevance of AI tools for managing human infections.

Main Methods:

• A comprehensive literature search of 1617 PubMed results was conducted, prioritizing clinical trials, systematic reviews, and meta-analyses.
• The review focused on studies utilizing prospectively collected real-world data with clinical validation and research with translational potential.

Main Results:

• AI demonstrates potential in laboratory diagnostics (e.g., antimicrobial resistance profiling), clinical imaging (e.g., tuberculosis diagnosis), and clinical decision support (e.g., sepsis prediction).
• Current studies often lack real-world validation and clinical utility metrics, with significant heterogeneity in design and reporting.
• Practical and ethical challenges, including algorithm transparency and bias, need addressing.

Conclusions:

• While AI development for infection research is rapidly advancing, its current real-world clinical utility is modest.
• Further research and validation are crucial to realize the full potential of AI in clinical infection practice.