Prof. Kamila Hawthorne from the Royal College of GPs stresses the importance of closely regulating AI to ensure patient safety and data security.

A recent study conducted in February 2024 surveyed 1,006 GPs in the UK, revealing that 20% are using artificial intelligence (AI) in their practice without formal guidance or regulation.

The findings highlight a growing reliance on AI in healthcare, particularly for documentation and clinical decision-making, with nearly one-third of GPs using AI tools for generating documentation after patient consultations.

Among those utilizing AI, 28% explored alternative diagnoses, and 25% relied on AI to suggest potential treatment options, raising concerns about patient safety.

Alarmingly, the fact that a quarter of GPs are using AI to propose treatment options underscores the urgent need for clearer operational policies regarding AI implementation.

Dr. Rob Hosking of the RACGP warns against using AI for clinical advice until these tools are validated against quality standards, emphasizing that current technology is too broad and risky.

While Dr. Hosking views the adoption of AI as beneficial for administrative tasks, he cautions against its use for clinical decision-making until further validation is achieved.

The research team emphasizes the necessity for medical professionals to understand the potential advantages and risks of AI, including inaccuracies, biases, and patient privacy concerns.

Concerns have been raised about the accuracy of AI-generated information, with medico-legal advisers warning that AI responses may contain inaccuracies or refer to incorrect guidelines.

Researchers highlight the need for compliance with regulations to protect patient privacy when using generative AI tools, particularly in light of how internet companies handle data.

The study concludes that while AI tools can aid GPs, they also carry limitations that could result in errors and privacy issues, necessitating further research on their safe and secure implementation.

The survey results, published in the BMJ Health & Care Informatics, indicate a significant trend towards AI integration in healthcare, yet underscore the need for better training for doctors regarding the benefits and risks of these technologies.