“Deploying an ethical artificial intelligence will be the challenge of the coming century”

AWhile the ChatGPT phenomenon – a conversational application capable of providing small texts at the request of users – impresses experts and laymen alike, the possibilities that this technology opens up force a rethink of the ethical question in artificial intelligence (AI) , so that progress takes the right path, at the service of the greatest number.

In his Ethics in NicomaqueAristotle (384-322 BC) conceived of ethics as an individual necessity, making it possible to achieve happiness through the prism of virtue, but also as a collective imperative, necessary to contribute to a more fair.

Now imagine Aristotle back among us, discovering artificial intelligence and questioning its ethics. What would he say? No doubt that AI is one of the greatest technological feats ever achieved, that it irrigates the whole social and economic world, but that, alas, the notion of ethics is largely absent.

At first glance, AI and ethics seem irreconcilable. Ethics proceeds by trial and error, approaches moral cases in their complex singularity, does not bother with absolute rules. Artificial intelligence, on the contrary, is based on formal and binary scripts; it rarely deals with nuances.

AI with positive and democratic impact

Thus, online advertisements for baby products continue to target women who have had a miscarriage, at the risk of totally traumatizing them. Stopping targeting them wouldn’t be difficult though: it would just require a bit more programming and attention. Ethics and artificial intelligence cannot therefore be opposed: an algorithm can be ethical; it is generally sufficient that it be built with ethical concerns in mind.

Read also: Article reserved for our subscribers David Bounie and Winston Maxwell: “Integrating ethics into algorithms raises titanic challenges”

What then could be the cardinal principles of an ethical AI? An ethical AI is first and foremost an AI with a positive impact on the environment, the economy and society in general. It is also an AI capable of operating without discriminatory biases – unlike, for example, mortgage allocation algorithms, which expert Aaron Klein shows that they structurally favor white applicants in the United States.

An ethical AI is a democratic AI, that is to say understood and accessible by all. The confusion surrounding Parcoursup’s algorithms shows how urgent it is for citizens – even the youngest – to understand and be able to audit AI, as it increasingly presides over individual destinies.

You have 47.54% of this article left to read. The following is for subscribers only.