Yesterday I presented the first of two webinars for the FAIMER1 Connect Learning series, entitled, Coming to terms with generative AI. The second session will be in November, focused on education, research, and clinical practice.
In this session I covered the main concepts related to generative AI, its implications for professional practice, and discussed some of the potential for the future in health professions educational contexts.
Here are the detailed key points I cover in the presentation:
Understanding generative AI
- Generative AI, like ChatGPT, Claude, and Gemini, is not just a next-word predictor but a multimodal system capable of handling text, audio, images, and video.
- It’s rapidly improving and becoming ubiquitous in our daily lives, from operating systems to cars and phones.
- Unlike traditional search engines, generative AI doesn’t retrieve information from a database but generates responses based on patterns in its training data.
Implications and challenges
- The scale and pace of AI development are outstripping our ability to regulate it.
- AI provides widespread access to deep expertise across various domains.
- It’s a tool that can be used to create new tools, potentially leading to exponential growth in capabilities.
- There are concerns about copyright, environmental impact, equity, and cultural influences, though these are debated.
Effective use of AI
- Prompting is important: It’s about establishing the context for the AI’s response, not just keyword searching.
- Structured prompts with clear roles, goals, and instructions yield better results.
- AI can be used as a writing assistant, for literature reviews, and even for creating simple websites or course outlines.
AI in practice
- AI can significantly reduce skill gaps between top and bottom performers in various domains.
- The combination of human expertise and AI capabilities often outperforms either alone.
- AI is particularly useful for tasks like summarizing documents, providing analysis in specific contexts, and generating ideas.
Challenges and future directions
- Many people struggle to take full advantage of AI due to low digital literacy or lack of institutional support.
- There’s a need to integrate AI into existing systems and workflows, which may require broader systemic changes.
- As AI becomes more advanced, we may see networks of AI systems acting as autonomous agents across society.
- FAIMER is the Foundation for the Advancement of Medical Education and Research. I was a FAIMER Fellow and Faculty member in the past, completing their Fellowship programme in 2010 (see my Related Posts on FAIMER). ↩︎