Article 22 of the GDPR says: “The data subject shall have the right not to be subject to a decision based solely on automated processing, including profiling, which produces legal effects concerning him or her or similarly significantly affects him or her.”
I used to think that this GDPR article is impossible to implement because as a data scientist , how can you guarantee that whatever decision pops out of such a process is not based “solely” on automated processing. I think this controversial article for a set of specific use cases including decisions impacting the health of individuals actually guarantees the indispensable human oversight in any AI solution.
I have been active in the health tech sphere for quite a while, and I am a firm believer in the complementarity between health and technology. Improving public health is a major motivation factor for technological experts and technological innovation pushes medical progress forward. Technology can automate, standardize and anticipate.
But I am also convinced that human interaction should be at the basis of health tech projects facing patients I have read a lot about a concept that arose a few years ago: humanized technology. It has almost turned into an obsession for me and a constant point of attention in my daily work.
I could summarize it in one sentence: nothing should ever supersede the patient-doctor interaction. In a way, I want to go even further than humanized technology (for which the keyword is still technology…),
I want to apply technological humanization. Make the world more humane thanks to technology.Practically, technology should be used to enable a doctor to get in touch, to spread information, to talk to patients as often as it is needed. It should also alleviate the administrative burden of doctors so that they have more time to discuss, to listen, to explain treatments and options to patients.
Technology should also empower the patient, so that the medical relationship they cocreate remains a level playing field. In a nutshell, my conviction is that technology is a faithful slave, but a lousy master. That is why it should be used to boost human relations, not to replace them.
I find it wonderful that when a ML software detects a tumor, a radiologist has to confirm it before any action can be taken. Human intervention always comes first. It is also a way to acknowledge that AI applications have been elaborated by human beings, and thus are never flawless. In spite of all tech fantasies, technology is not neutral, algorithms can even be biased. Leaving the final say to a human being reduces the negative consequences of potential flaws.
How do you envisage and apply technological humanization?