“Without big data, you are blind and deaf and in the middle of a freeway.”

Geoffrey Moore, American author and consultant

Large language models (LLMs) like that seen in ChatGPT or AutoGPT use deep learning and natural language processing, and this AI resource will impact on many aspects of healthcare, including research. In addition to data analysis and predictive analytics, these language models can be useful as a natural language tool for literature review and manuscript summaries as well as electronic record information extraction.

Natural language processing (NLP) is how machines communicate with humans and consists of two components: natural language generation and understanding. Common natural language processing tools include: 

  • Lemmatization
  • Named entity recognition (NER)
  • Part-of-speech tagging (POS)
  • Parsing
  • Segmentation
  • Stemming
  • Tokenization (or word segmentation)
  • Word sense disambiguation

We are witnessing a Cambrian explosion of NLP tools with an impact on healthcare.

Graph database is a database that is potentially better suited for the complexity and heterogeneity of biomedical and healthcare data. Instead of traditional tables with rows and columns with relational databases, graph databases are more representative of complex relationships and use node and edge connections to represent these relationships. This graph database is also convenient for knowledge graphs, which can be very useful in biomedicine and healthcare.

While successes of AI in healthcare this decade have included medical image interpretation, protein structure determination, and healthcare administration (with robotic process automation), the myriad of relative disappointments include lack of continual access to high-quality healthcare data, poor performance of some decision support tools, and the ongoing schism between projects/publications and actual AI in clinical practice or healthcare administration.

The future of clinical research may involve a healthcare learning system (brain) with real-time data and deep reinforcement learning, and be intertwined with a digital twin concept of patients. The model in the future may be coupled with strong cognitive elements to achieve true human-machine synergy. Patient care and clinical research may be much more intertwined as a new paradigm via the use of advanced AI tools such as deep reinforcement learning and cognitive architectures.

Finally, a supremely important dimension of AI in healthcare is the constellation of “REAL” issues:

R for regulatory

E for ethical

A for accountability (for health equity)

L for legal

The technological development of AI has been on an exponential trajectory, but these “REAL” issues have not kept up with the frenetic pace of the technological development of AI. These issues are increasingly intertwined and complex, thus demanding “human”, not artificial, wisdom.

The importance of these observations of artificial intelligence in healthcare will be part of the topics of discussion at the in-person AIMed Global Summit 2023 scheduled for June 4-7th of 2023 in San Diego. The remainder of the week will be other exciting AI in medicine events like the Stanford AIMI Symposium on June 8th. Book your place now!  

We at AIMed believe in changing healthcare one connection at a time. If you are interested in discussing the contents of this article or connecting, please drop me a line – [email protected]