A Year In AI: The Most Important 2023 Milestones For Medicine

In 2023, the widespread adoption of generative AI, notably OpenAI's ChatGPT, marked a transformative year in healthcare AI. The public access to ChatGPT and its success, with over 100 million users in two months, showcased the potential for broad medical applications. The release of GPT-4 and exploration of multimodal capabilities further fueled healthcare innovation but underscored the need for comprehensive regulatory and privacy frameworks.

The surge in AI development prompted the creation of specialized models like Med-PaLM by Google/DeepMind, specifically designed for healthcare-related questions. The release of Med-PaLM 2 demonstrated impressive accuracy in answering medically related queries, albeit with room for improvement. Regulatory issues became a prominent concern as AI capabilities expanded beyond text-based interactions to analyzing various data types, necessitating a flexible framework for regulation.

Medical professionals faced the challenge of integrating AI into healthcare responsibly, emphasizing the need for a basic yet robust understanding of AI's safe and effective application. Regulatory oversight of large language models (LLMs) in healthcare became imperative, as discussed in a paper by The Medical Futurist and Dr. Eric Topol.

Privacy concerns and the multifaceted challenges of AI in healthcare spurred discussions about regulation and ethical use. Despite these challenges, a growing understanding of generative AI became crucial for healthcare professionals, fostering a high-level understanding of its applications.

The year witnessed the predicted future becoming a reality, with developments such as multimodality becoming present in AI models like ChatGPT. OpenAI integrated various modes into a unified user interface, allowing users to upload visual prompts. Google's announcement of Gemini, a natively multimodal algorithm, further exemplified the evolving landscape of AI.

Amid doomsday scenarios and media hysteria, a positive attitude prevailed, backed by optimism and the recognition of AI's potential. While acknowledging the need to address regulatory, ethical, and moral questions, the year concluded with excitement about the birth of multimodality and anticipation for further advancements in 2024.