Can artificial intelligence really help science to make diseases disappear? This idea, which is strange at first sight, will undoubtedly make skeptics of modern technology smile. Scientists, on the other hand, see the potential and several solutions are already being implemented.
Altruistic? Maybe. Useful? No doubt about it.
Our world of technophiles is sometimes pushed to the extreme: many fans will defend the brands they adore to a point that even worries some sociologists. This is the power of marketing in its raw state, and it’s neither new (as this 1994 article attests) nor truly surprising since current generations were born in it and are bathing in it without even realizing it.
This is a worrying subject and techno-skeptics tend to rush into the same arguments: “technology companies are interested in nothing but themselves”, “they don’t want to help users as they claim, they just alienate them to keep them in their midst”, etc. These points, which are generally well founded, still need to be qualified. Many well-known technology companies are opening up new scientific horizons, and while their interest in medicine isn’t entirely altruistic, it is nonetheless real and can make a difference.
A development where medicine finds its place
Artificial intelligence is above all a technology, so it isn’t surprising to see tech giants looking into the subject. Amazon launched Division 1492, a department dedicated to connected health that works with machine learning. If you didn’t get the gist of the name, Amazon considers the field of digital health as a new continent to explore, just like Christopher Columbus did in 1492. Google / Alphabet is of course also looking at these technologies, and we’ve already discussed their efforts in the field. Facebook, on the other hand, is also ambitious and wants to eradicate all diseases by 2100. Microsoft has even bigger ideas: it plans to beat cancer by 2026.
Beyond the (probably optimistic) ambitions of these companies, it is interesting to try to understand how AI could revolutionize medicine.
Artificial intelligence has evolved enormously in recent years. Remember in 1997 when astrophysicist Piet Hut said that it would take a hundred years before a machine could beat a human at the game of Go? In 2017, the AlphaGo program created by DeepMind (a Google subsidiary) managed to beat the world Go champion into the ground.
The evolution of artificial intelligence goes much further than just a game. The huge database it needs to anticipate all possible scenarios can now feed itself thanks to the concept of machine learning. The more data AI has, the more effective it can be. It's like a human, in short, because its ability to react will depend on its knowledge, but the machine is much faster.
What uses does AI have?
The computing speed and the precision with which a machine processes lots of data can be a crucial asset in medicine. However, it isn’t intended to replace a doctor’s diagnosis; it’s meant to complement it.
Laurent Schlosser, director of Microsoft’s public sector division, explained that: “98% of health today is curative. AI will make it possible to switch to a more preventative medicine”. He’s implying here that the strength of AI lies more in the early detection of problems than in their treatment.
He’s undoubtedly most interested in cancer, a real modern plague. The understanding and detection capabilities of AI allow decisions to be made that in combination with a medical decision, can lead to the detection of diseases at an early stage. That means oncology will likely be a field where AI is applied in medicine since detection plays such an important role. Artificial intelligence has been developed in China that is sometimes better at detecting tumors than a team of 15 doctors. We should also mention Google’s retinal analysis, which with a simple scan of the eye, provides a wealth of information on the patients' health and can anticipate possible strokes.
DeepMind, which created the formidable AI for the game of Go, has a very noble goal: to use AI to reduce the number of deaths in hospitals by limiting human errors made by doctors. By accessing patient data, AI would be able to determine whether patients will have a problem or not by analyzing all accumulated data (last operation, history, age, etc.).
98% of health today is curative. AI will make it possible to switch to a more preventative medicine.
Many or other companies, both big and small, are looking into the topic. Several solutions are already available, generally at a less specialized level. We should mention Babylon Health, an online consultation app. The London-based company is working hard on artificial intelligence to obtain diagnoses, and the NHS is convinced that it’s being tested as an alternative to NHS 111 (the helpline number dedicated to medical emergencies in England). Its success rate is reported to be 92%, which is more impressive than doctors (82%) or nurses (77%).
Vigilance is required
As is always the case with technology, we always come back to the same question: will machines eventually replace us? This is certainly possible, but it won’t happen in the near future. The medical technologies associated with artificial intelligence aim to understand problems, detect anomalies, and perhaps analyze solutions. They create a complementary diagnosis that aims to see what medicine doesn’t see itself. The future will show us how this situation is evolving.
There’s another problem that’s a little more alarming: digital companies are taking control of new markets. Will we ever see Google hospitals? Facebook clinics? Will patients automatically be assigned an advertising ID?