There are four distinct terms that are important: artificial intelligence, machine learning, big data, and robotics.
According to the Oxford English Dictionary, artificial intelligence describes “the study and development of computer systems that can copy intelligent human behavior”; machine learning “a type of artificial intelligence in which computers use huge amounts of data to learn how to do tasks rather than being programed to do them.” Big data is defined as “sets of information that are too large or too complex to handle, analyze or use with standard methods.” Finally, a robot is considered as “a machine that can perform a complicated series of tasks by itself.”
Once reading the definitions, one understands that artificial intelligence is the overall entity that researchers have used to build robots, and that machine learning enables the robot to “evolve” itself, adapting and adjusting to perform better using its own experiences. In a recent review article on artificial intelligence,3 the importance of input by practicing clinicians in the further development of devices using artificial intelligence was pointed out: one could imagine that future robots will be put into use and that they might “improve” on the job, as humans should do, using machine learning based on experiences, feedback from the clinicians, and program changes. Big data will be helpful because the more “data” are available, the more the robot can use to improve: big data are the computer equivalent of years of experience of a human anesthesiologist.
Obviously, robots improving on their own using machine learning will be a significant challenge not only for the developers but also for regulatory entities, which will have to “recertify” or “re-evaluate” these robots. It is also important to notice that machine learning is still in its infancy with some research in anesthesia showing promising results. Arora1 states that machine learning algorithms are capable to outperform human decision-making provided that their data set is reliable. I personally believe that machine learning will “take off” with the widespread creation of big data and its use to develop appropriate machine learning algorithms. The more data can be fed into a machine learning system, the better this system can adapt. To repeat the former analogy, the more experience an anesthesiologist has, the more he or she has experienced, and his or her techniques have been able to evolve and change. I believe that future developmental strategies should adopt a concept where new big dataare constantly fed into a robot in order for it to improve itself, similar to anesthesiologists who grow wiser and better with years of training, training in simulation, reading books and articles… This concept could be a sort of “continuous medical education” for robots!
The author declares no competing interests.