By Dr. Colin Michie FRCPCH
Ah...
We discover, we ourselves
Are the Amazement which awaits
We are ourselves Amazement.
- Maya Angelou
The Parliament of St. Maarten promised to explore a national artificial intelligence strategy in 2026. Artificial intelligence (AI) will change our lives and our health.
The label “artificial intelligence” was used in 1955; thirty years later a robotic arm was employed in a surgical theatre. Learning machines and robots are now everywhere. You probably already use chatbots to answer your questions, perhaps to write letters, deal with your homework or analyse your gym data. Using AI is already much like using electricity.
AI processes have mind-blowing speeds and capacity. Reviewing documents, spreadsheets, images and sounds takes place in the twinkling of an eye. Last year one study documented AI checks of over 460,000 mammograms collected from women attending 12 American clinics. It delivered a higher rate of cancer pickup, lower rate of recalls and more precise diagnoses than routine screening. This means future scan workups can be improved across wide geographical areas.
In a smaller investigation, a trained AI mental health chatbot provided talking therapy to 210 patients over four weeks. The “bot” engaged with the patients and established effective therapeutic alliances. Patient scores for depression and anxiety improved, while waiting lists were shortened; staff spent more time working as healers.
Surgical robots race ahead as particularly effective, practical enablers for surgeons. From the brain to the toes, their operations are less invasive, more precise and controlled than before. Tumours, bony masses or obstructions are removed while using imaging during surgery. This improves outcomes! Soft, flexible robots have been built that apply techniques from origami to fix and hold themselves in useful positions while minimising tissue damage. The miraculous mechanisms of an elephant’s trunk have been copied to help robotic gripping. Surgical robots probably represent the ultimate industrial revolution (level 6), because some of them have assembled their own industry: intelligent manufacturing. This builds personalised robots alongside surgeons for specific patients, using AI and advanced technologies.
Scientists use AI’s skills and power to open new horizons in their research. The 2024 Nobel prize for chemistry was awarded to the researchers who developed an AI tool, AlphaFold. This predicts the exact shapes of molecules such as proteins or DNA. AlphaFold is freely available and is steadily improving: it is now in its third generation and used by millions of researchers, including many from low- and middle-income countries. It guides the design of medications and vaccines rapidly and at lower cost, replacing processes that once took many years in expensive laboratories. This approach has other benefits – it reduces the number of animal experiments, such as those using primates. AlphaFold allows exploration of genetic mutations too, in humans, livestock, crops, parasites or microorganisms, to search and find methods to tackle diseases more effectively in future.
ChatGPT, a large language model, has been trained using vast amounts of text from the internet. Computing predicts its next word so swiftly it can hold conversations or translations in real time. Such systems do not think, or have ideas; rather they link words from their massive memory banks. They remember more efficiently than we can too. Working with a ChatGPT ‘bot’ or chatbot as a sidekick, friend or wise health coach can be positive. It may streamline your diary, promote your confidence and give advice for your health, beauty and lifestyle (as well as a recommended playlist). Despite all of this power, they are not human!
AI sometimes gives rise to problems in healthcare because of their training. If language or data are not on the internet, AI systems are unlikely to ‘know’ about them. Cultures, castes, rural or island communities, or parts of the global south are poorly represented on the internet. AI may therefore have biases. Some AI on testing has been found to have biases against women too, because of the materials used in their training.
AI’s ethical guardrails in healthcare may need checking. Are patients and their families safe, their information confidential? Are recommend treatments the best available? Is an AI system transparent, accountable, trustworthy? Is there human oversight and respect for patient autonomy? Advertising can prioritise what we are shown as available for our healthcare because of commercial drivers. Problems can also develop if AI devices cannot answer a question: they may make up a response with a misleading fake or “hallucination”. Human discussions and backups are vital!
AI will help with challenges from climate or social changes. It will be invaluable in tackling new microbes, resistance to antibiotics, wearable devices. But it has to be ethical. Our education and common sense means we can check superhuman AI systems. Can they benefit human welfare safely to become amazing, virtual Best Friends Forever?!
Dr. Colin Michie specializes in paediatrics, nutrition, and immunology. Michie has worked in the UK, southern Africa and Ghaza as a paediatrician and educator and was the associate Academic Dean for the American University of the Caribbean Medical School in St. Maarten a few years ago.
Useful resources: deepmind.google/science/alphafold ~ digital.nhs.uk/services/ai-knowledge-repository.
