What AI actually is — and is not
When healthcare professionals hear "AI," many think of diagnostic imaging algorithms or clinical decision support tools — the kind of AI that reads an MRI and flags a tumour. That is a real and important category of AI, but it is not what this course is about.
The AI we are covering — large language models (LLMs) like Claude, GPT-4, and similar tools — is fundamentally a language processing technology. It reads text, understands context, follows instructions, and generates text. Think of it as an extraordinarily capable research assistant that can read, summarise, draft, compare, and extract information from documents at scale.
What makes this relevant to healthcare and pharma is that so much of the operational work is text-based: clinical notes, discharge summaries, prior authorisation letters, clinical trial protocols, regulatory submissions, adverse event reports, coding queries, patient communications, and compliance documentation.
An LLM does not "know" medicine in the way a physician knows medicine. It does not have clinical judgment, it cannot examine a patient, and it does not have access to real-time clinical data unless you explicitly provide it. What it can do is process, summarise, draft, and cross-reference text-based documents faster and more consistently than any human.
