Table 1.

Advantages/Potentials and Disadvantages/Pitfalls of AI and LLMs in Managing Superficial Fungal Infections

Potentials	Pitfalls
Pathology Settings	Model Accuracy
Increased throughput of samples	Can be prone to hallucinations/confabulations
AI assisted analysis of samples	Quality of training data can impact accuracy
Report generation in genetic testing for species identification and detection of resistance mutations	Falsehood mimicry due to accepting false information from user as fact
	Algorithmic errors
Patient Screening	User experience in prompting
Differentiate infectious vs non-infectious conditions	Potential for patient harm
Reduce need for unnecessary testing	Stale-dating of training data
Reduce the need for unnecessary antifungal prescriptions	Data bias
Promote anti-microbial stewardship
	Regulatory and Ethical Concerns
	Breach of patient privacy
	Unauthorized use of patient information
	Regulatory status and implications
Access to Dermatological Care	Process transparency
Consultation tool to assist non-dermatologists in providing care
Formation of treatment plans personalized to each patient	Legal Concerns
	Who is responsible party in event of medical error
	Access Concerns
	Cost of specialized hardware
	Cost of running the model
	Availability of sufficient infrastructure
	Adequate training for users