Advancements in AI make headlines in the press, become a topic among friends, become topics in the classroom — as they say, conversation starters. Part of this fascination comes, of course, from all the imagination that has been created around the concept. And this fantasy is built, in turn, on the relative lack of knowledge that still exists about what intelligence is. Incidentally, an excellent read on the subject is now the classic “Structures of the Mind – The Theory of Multiple Intelligences,” by psychologist Howard Gardner.
Another facet of AI that’s generally ignored — but you can’t pretend it doesn’t exist — is responsibility. We have, with relative generosity, handed over some of our responsibilities to AI. We schedule payments in Finance apps. We give access to our phone’s photo and contact files to some other app for it to work. We entrust certain tasks to algorithms, freeing up time to devote to other things. As they get more complex, these applications and algorithms, and other systems in which AI operates, work just as well as we expect them to. A small part of our daily lives only happens relatively smoothly because AI works so well. Until something goes wrong.
Efforts to make AI make fewer and fewer errors continue. There are already algorithms that try to prevent traffic accidents, work accidents in industries and civil construction, and even in the health field – by making the diagnosis more accurate. Machine learning helps reduce the need for human intervention – and thus reduces the risk of accidents due to the so-called human factor.
But AI errors in facial recognition, for example, have happened. In the United States, since 2020, there have been cases of people getting caught for errors like this. In the same year, the UK education system suffered a wave of grading errors that left everyone scratching their heads – the main reason behind this was artificial intelligence.
Such errors cause tangible material harm to people. The lives of those caught by an AI bug could be turned upside down, with consequences that may never be fully reversed. Let’s also think about the mistakes that could ultimately affect banking systems – and put our financial information at risk.
One of the first questions – if not the first – to be asked in these contexts is: Who will be held accountable? Victims of these situations can face a long and grueling legal battle, which can end without the damage being remedied proportionately. Legislation to deal with liability for errors made by artificial intelligence is still in its infancy. In Brazil, we already have the LGPD (General Personal Data Protection Law), which dates back to 2018. Your role is to protect personal data and the privacy of each, by digital or physical means. But the nature of technological development will lead to the constant modernization of legal provisions – at the risk of becoming obsolete in a short period of time.
The United States is surprisingly behind in this regard. Just last month, the OSTP (Agency for Science and Technology Policy) introduced the Bill of Rights for Artificial Intelligence. In Europe, the authorities are closely monitoring the performance of the tech giants with regard to the destination given for information shared by users online. The difficulty of assigning responsibilities is verified by the many times that the owners or key executives of these companies are called to testify before parliamentary committees, due to the legal battles in which they are involved.
Digitization is a process that permeates everything and, as far as we can see, is not going to stop. As AI will be more and more present and, despite its increasing effectiveness, will never be fault-proof, the possibility of harming a person increases. The technical framework needs to continue to evolve so that the margin for error is smaller and smaller. The legal framework for accountability for wrongdoing must be in place and up to date. Not only that: inspection and enforcement of penalties must act – otherwise the laws will be mere paper tigers.
MD, he is Chairman of the Board of Directors of the Institute of Radiology (InRad) and Chairman of the Innovation Council (InovaHC), both at the University of the South Pacific School of Medicine’s Das Clínicas Hospital, and Vice President of the Kolejão Institute. Said (Equus).
“Friendly zombie guru. Avid pop culture scholar. Freelance travel geek. Wannabe troublemaker. Coffee specialist.”
