Are ethics and artificial intelligence compatible?
I already had received Aurélie at the very beginning of this podcast moreover on a more general subject around AI.
And that is precisely why I asked Aurélie to come back and talk to us about a subject that seems essential to me in the 21st century, namely ethics.
It’s a bit of a special episode since it was recorded live with an audience at MylittleParis as part of their summer school.
I will have the opportunity to record others with the public so especially write to me if you are interested.
Are ethics and artificial intelligence compatible and if so how?
In Europe and a fortiori in France, we have a strong posture with regard to ethics and it is therefore necessary to look into this strategic subject.
Yes Artificial intelligence and ethics can work together, asserts Aurélie.
If we take the 3 models in the world around AI and ethics, we have on the one hand and very schematically the US who consider data as a commercial element (even if Aurélie reminds us that everything is not so liberal in the country of Uncle Sam), China which considers data as an element of power in order to define “good citizens” and Europe in which, with the GDPR recently, we are trying to protect data.
Regardless, she is convinced that ethics and artificial intelligence are not at all opposed.
In reality, ethics is never defined in “good” or “bad” but rather in point of view obviously.
Simply according to values, ethics and morals can evolve from one country to another.
According to Aurélie, the GDPR is also a big step in this direction and explains how the GAFA are obliged to respect it at least for Europe.
It will be relevant to look at how this first version impacts uses in order to change the law accordingly.
What is certain is that we must educate because for the first time our leaders do not understand the revolution in progress even if they try to be interested in it.
How cognitive biases influence the ethics of artificial intelligence?
Every human has cognitive biases, that is to say that his way of perceiving the world is necessarily different from another and this will be found in artificial intelligence.
As Aurélie points out, there are well-known examples such as the inability of smartphones (for the 1st models) to offer facial recognition that works for people of color but she also talks about her own subjects of studies on trauma cranial and therefore the size of human skulls on which she relied in her work.
She then went on a large majority of skulls of white people until she realized that Asian people had different skull sizes. This is a simple example but we could make a similar mistake.
In short, even with the best will in the world, biases are everywhere.
And obviously, this ends up creating algorithmic biases that we find in machines but it is not a desire not to be ethical or to harm on the part of scientists, you will understand that.
According to Aurélie, it is the responsibility of all the people working on a project to be aware of cognitive biases because it is the only way to stop their propagations.
It is up to everyone to become a little more philosopher in order to move in this direction and especially to promote transdisciplinarity in order to be more relevant.
Aurélie worked with Grégory Renard on an artificial intelligence oath to avoid errors in judgments.
It can be found online in French and English and aims to defend life (not just those of humans of course).
It includes data protection, peer protection and the transmission of information …
Can we be competitive and ethical?
It’s interesting that even when the boundaries of ethics are widened it doesn’t necessarily solve everything.
Aurélie takes as such the example of human cloning which, even if it is potentially accepted in China, has still not taken place.
But most importantly, ethics is a weapon of consumers.
Consumers are the Achilles heel of artificial intelligence and tomorrow Aurélie envisions that consumers will ask to use products and services with an ethical label that will respect their privacy and their data.
Perhaps even consumers will be willing to pay a little more to have their data used only for service improvement and not for commercial purposes.
When Youtube offers to pay to access an ad-free space, it is not the same promise. We make it easier for you to use it, but your data is still captured. Ditto for Netflix or Spotify.
For Aurélie, Europe therefore has a competitive card to play by taking this ethical stance, of services that are just as effective but respectful.
Even Europe could become a land of welcome for people who want to protect their data.
Gregory Pouy is the founder of LaMercatique, a digital transformation consulting firm focused on the marketing part. Based between New York and Paris, he is a marketing “expert” for FrenchWeb.fr. To follow his writings and exchange with him