- Interview -
Is our human authenticity
under threat from artificial intelligence?
Artificial intelligence is enjoying a spectacular surge, daily reconfiguring every aspect of our lives. It’s a dizzying prospect that also poses a challenge to society and humankind in general. A conversation with Jean-Gabriel Ganascia, author of Le mythe de la singularité. Faut-il craindre l’intelligence artificielle ? (The Myth of Singularity. Do we have anything to fear from artificial intelligence?) published by Seuil.
What led you to take an interest in artificial intelligence in the 1990s already, at a time when it was not yet a particularly hot topic?
Having studied both engineering and philosophy, I found the idea of working on analyzing thought using machines was an excellent means of combining my two fields of interest. I also did my military service with a unit dedicated to studying nocturnal surveillance images. It was such a painstaking exercise that I wondered whether these recognition systems could be automated so as to avoid human beings having to spend whole nights in front of screens. All of which converged towards the idea of artificial intelligence.
These days AI is everywhere. Will it overturn our societal points of reference?
Our landmarks are constantly shifting. Our social fabric is governed by a certain number of fundamentals that evolve in step with society. One of them is friendship. It has always existed, but can the kind of friendship that existed in Ancient Times be defined in the same terms as that we experience in the era of social media? While it continues to be expressed through particular affinities between individuals which drive them to engage in contacts with each other, is the feeling itself the same? Is it as authentic today?
Is our authenticity as human beings declining?
I think that the reason artificial intelligence and authenticity are often regarded as polar opposites, is that the intelligence of machines is considered non-authentic, contrary to that of humans. But this issue deserves deeper thought, because artificial intelligence is not really that of machines. It is instead a scientific discipline that attempts to simulate human intelligence via machines. The latter are not intelligent in and of themselves, but rather thanks to the intelligence with which we as human beings endow them.
And they are changing our lives…
They are doing so in many respects. Taking only the example of information systems, they offer us continuous access to an infinite amount of data. This is very exciting and at the same time very worrying, because we are left to our own devices to handle this mass of information, whereas previously we were given texts to read by our teachers who also taught us to analyze them. The issue is this: will machines at some point be able to play this role, namely guiding us by means of their extremely broad observation base? Would that be a satisfactory solution? And will we be able to retain our freedom to decide in the face of these engines that are better informed and more experienced than us? There are a lot of questions to be answered…
As far as access to knowledge is concerned, it seems that we may be able to implant extra memory in our brains. Is that really conceivable?
This is indeed part of a venture undertaken by American businessman Elon Musk, who also thinks that computers will soon be more intelligent than we are and will seize power from humans. Based on his reasoning, in order to vie with machines, we will need to reinforce our cognitive capacities and boost those of our brains, notably in terms of memorization – hence the implant project. Personally, I think our gray cells harbor far too many mysteries to envisage this. And our memory, which is extremely complex, cannot be summed up in terms of mere information storage. Especially since it is not located in just one area, but is instead spread throughout and can be divided into several categories: subconscious, conscious, associative. Finally, even if it were possible, it would be so terrifying that it should be avoided at all costs – since the master of the device could implant anything he wanted into our brain, or even erase everything if he so wished!
You have a longstanding interest in visual recognition. What are the applications for this technology?
There are hundreds, some of them very positive. For example, in the field of preventive medicine, certain skin cancers can be diagnosed based on photos of moles. Others are equally fascinating but also debatable. A machine can recognize millions of faces with 99% precision, which is far higher than a human being can achieve. We could therefore choose to equip police forces with devices for detecting individuals, place these technologies on traffic lights, or slap fines on anyone stepping across the road outside pedestrian crossings… Although this might reduce the number of accidents, it cannot be undertaken lightly. These subjects must be the object of ethical questioning, if we wish to leverage the considerable progress afforded by machines, yet without transforming our society into a living nightmare.
Are machines capable of creativity?
There is something magical about the idea of ‘creation’: a certain sensitivity and soulfulness that physics cannot begin to grasp. Machines are not going to start imagining things or writing great novels, but they can simulate certain aspects of the creative process and help us understand them better. By way of example, at one stage we decided to work on jazz improvisation. The idea was to start with a model, to pinpoint the musical elements associated with this context, and then combine them according to certain laws. This generated some very interesting results, even according to the musicians who commented on the experiment.
I don’t believe in a theory of any sudden and dramatic pivotal moment when machines would seize power and overturn the status of humankind forever. I do not see this ancestral myth of the creature turning against the creator as being anything we need to fear. However, it is certainly realistic to be wary of certain dangers, notably on a political level – given that compared with democratic states, rooted in specific territories and governed by rigorous rules, the digital field affords infinite space and power to forces that have no need to account for their actions.
Towards what model of society is artificial intelligence driving us?
Towards the one we are currently experiencing, but taken to extremes: namely a digital society, which began developing gradually in the 1960s, before a fierce acceleration in the late 1990s, with the development of inter-human exchanges. Within this society that has been taking shape for the past 25 years, artificial intelligence already plays a crucial role. It uses the vast mass of information to build new knowledge, such as we see with Big Data. It creates search engines to help find a way around the vast body of data; it develops languages for communication with machines… It is already part of our daily life.
And will we be able to keep control?
I am convinced there is no determinism with regard to machines and that we will always maintain power over the decisions to be taken and the societies we wish to have. What will happen will depend on us and on what we make of it. We are free. Such is the destiny of humankind: to be able to make use of this freedom for better or for worse. This freedom makes us responsible for the world to come. If used badly, this technology can generate terrifying societies. If handled wisely, it opens up a huge and wonderful field of possibilities.
Based on an interview by Michèle Wouters
Professor and Researcher in the IT Lab of the Pierre and Marie Curie University, as well as President of the CNRS Ethics committee, Jean-Gabriel Ganascia has been exploring artificial intelligence for several decades.