Artificial intelligence in pharmacy: A guide for clinicians

Extract

While artificial intelligence (AI) has seen a massive increase in attention in the last year, it is not new to medicine. In fact, as of 2023, the Food and Drug Administration (FDA) had approved nearly 700 AI- or machine learning (ML)–enabled devices.¹ This highlights the broad nature of the definition of AI. Namely, it is the use of algorithms or models to perform tasks and exhibit behaviors such as learning, making decisions, and making predictions. Given this definition, it is not surprising that not every AI/ML-enabled device represents a landmark change in medicine. For example, the microbial identification system Vitek MS (Biomerieux), approved in 2013, is included on the list. While this is an important tool for clinical care, for clinicians reviewing results from this instrument, it may not be clear that this tool uses AI at all.

While AI may not be new to medicine, the recent growth in this area is immense; 80% of the aforementioned AI- or ML-enabled devices were approved in the last 5 years.¹ Further driving the attention on AI is the newest generative tools, including large language models (LLMs) and their incredible performance as chatbots. Lay press headlines and even articles from peer-reviewed medical journals ask broad questions and make sweeping assertions about the dangers, fears, and promise of AI in medicine. Some titles suggest that complex, nuanced issues have simple, binary answers: “ChatGPT: friend or foe?”;² “Uncertainty quantification: can we trust artificial intelligence in drug discovery?”;³ “Artificial intelligence: the future of medicine, or an overhyped and dangerous idea?”.⁴ This can leave readers at a loss, struggling to reconcile these wide-ranging claims.