en-de  Artificial intelligence: reflections on its real dangers Hard
Artificial intelligence: reflections on its real dangers.

The explosive progress of computer science and robotics is worrying. Several publications aim at identifying the real stakes of these new technologies.

Le Monde - THE WORLD OF BOOKS | 08.04.2017 with 15:13 | By Céline Henne (Collaborator of the "World of books").

Translated from the French. LE MONDE DES LIVRES | 08.04.2017 à 15h13 | Par Céline Henne.

And if the greatest threat to mankind was not an ecological disaster or the nuclear bomb, but artificial intelligence? The idea that humans may one day live with robots that have become powerfully intelligent or threatened by them has always fascinated the world of science fiction and is still making money in Hollywood, Think of Terminator, by James Cameron (1984), A.I. Artificial Intelligence, by Steven Spielberg (2001) or Ex Machina, by Alex Garland (2015). In recent years, however, it has gone beyond the domain of fiction and has been relayed by scientists and engineers of great renown: astrophysicist Stephen Hawking and billionaire Elon Musk, co-founder of PayPal and SpaceX, have shared with others their concerns about the risks that artificial intelligence causes to our species. For Ray Kurzweil, project manager at Google, the issue is even inevitable: the exponential development of technology signals the programmed end of humanity as we know it, and it is only a matter of decades ...

While a wave of publications on the subject has appeared; the works of Laurence Devillers, Des robots et des hommes. Mythes, fantasmes et réalités (Of Robots and Men, Myths, Fantasies and Realities), and Jean-Gabriel Ganascia, Le Mythe de la singularité (The Myth of singularity). Should we fear artificial intelligence? Both professors of computer science and specialists of the question, are placing themselves subtly against the current of this new trend. For them, it is a question of unraveling the facts of fiction, and of proposing a realistic description of artificial intelligence in order to identify the true stakes. Of course, one can almost feel a little disappointed with this return to reality - some of us prefer to listen trembling to experts predicting the end of Homo sapiens. But let us reassure ourselves: the issues raised by the authors and the technological advances of tomorrow still have enough to provoke many emotions, between wonder and chills.

A prophecy passes for a scientific prediction.

The essay by Jean-Gabriel Ganascia aims at exposing the faults contained in the discourse of the "disaster traders," in particular the announcers of the "singularity", that is to say the moment of rupture that would constitute the emergence of an artificial intelligence capable of improving itself, and which would mark, for the human era, the beginning of the end. He is worried about the great intellectual authority enjoyed by "technoprophets", who contribute to a confusion between genres: what is, at most, a prophecy or a "vast cosmic narrative," passes for a scientific prediction!

The author thus applies himself to showing the limits of their arguments, both empirically and epistemologically, and denounces the confusion between the concepts. These new prophets all predict the passage of "weak" artificial intelligence, the one that is developed at the moment, endowed with autonomy in the technical sense (the ability of a machine to make a decision alone), to a “strong” intelligence, endowed with consciousness and autonomy in the philosophical sense (in which the machine would give itself "the rules and the objectives of its behavior"). However, he said, the prediction of such a qualitative leap cannot be justified in the current development of research.

The greatest danger, according to the two researchers, of these pseudo-scientific myths is that they divert attention from the real stakes present and future. For Laurence Devillers, "the current intelligence of the machines is already very worrying, the announcements shattering on their supremacy must not mask the ethical problems already present". Jean-Gabriel Ganascia thinks that these prophecies lead mostly to disempowering, hiding behind a "unique scenario presented as fatal", which leaves no room for freedom or action for man.

"Social robotics at the service of man".

Laurence Devillers delivers an enlightening insight into how artificial intelligence will overturn our ways of life and social interactions. In the short fiction that opens her essay, no exterminator robot or consciousness, downloaded on a computer, but sympathetic machines that act as aids at home and accompany people on a daily basis. Able to dialogue and simulate emotions, they even manipulate humor. This "social robotics in the service of the human being", on which her research team is currently working, carries many promises: the first versions of these machines have already proved themselves to people with Alzheimer's disease and autistic children.

Far from being optimistic, Laurence Devillers underlines the many ethical problems posed by these technological developments. How to manage the risk of too strong an attachment, or even dependence on these benevolent robots, in vulnerable people? How to ensure the data protection of these machines that will share our privacy? Who is responsible for the actions of a stand-alone robot and its "learner," its designer or its user? It is essential to provide for regulations on their use and regular monitoring, as it is necessary to implement ethical rules in their programs: both technical and legal challenges for tomorrow.

The economic imperatives of the digital giants.

But the main stake remains political. On this question, it is clear that it is the digital giants, the GAFA (Google, Amazon, Facebook, Apple) and a few others who seem to be leading the dance against increasingly powerless states. In the conclusion of his essay, Jean-Gabriel Ganascia accuses these multinationals of practicing a "bewitched charity": under the guise of warning about the dangers of artificial intelligence, they give themselves an image of benevolent enterprise, which can only inflect for the good of all a technological development which, in any case, is inevitable.

According to him, this strategy does not only respond to economic imperatives. Multinationals gradually take over the state's security, biometrics and health functions, which they pretend to assume better and at lower cost. They are creating a new society, the evolution of which they alone can master. So it is not the machines themselves that are to be feared and to be watched, but those who decide their ends and the rules of the game. One can trust man: he knows how to be frightened by himself.
unit 1
Artificial intelligence: reflections on its real dangers.
0 Translations, 0 Upvotes, Last Activity None
unit 2
The explosive progress of computer science and robotics is worrying.
0 Translations, 0 Upvotes, Last Activity None
unit 5
Translated from the French.
0 Translations, 0 Upvotes, Last Activity None
unit 6
LE MONDE DES LIVRES | 08.04.2017 à 15h13 | Par Céline Henne.
0 Translations, 0 Upvotes, Last Activity None
unit 14
Should we fear artificial intelligence?
0 Translations, 0 Upvotes, Last Activity None
unit 19
A prophecy passes for a scientific prediction.
0 Translations, 0 Upvotes, Last Activity None
unit 28
"Social robotics at the service of man".
0 Translations, 0 Upvotes, Last Activity None
unit 31
Able to dialogue and simulate emotions, they even manipulate humor.
0 Translations, 0 Upvotes, Last Activity None
unit 38
The economic imperatives of the digital giants.
0 Translations, 0 Upvotes, Last Activity None
unit 39
But the main stake remains political.
0 Translations, 0 Upvotes, Last Activity None
unit 46
One can trust man: he knows how to be frightened by himself.
0 Translations, 0 Upvotes, Last Activity None

Artificial intelligence: reflections on its real dangers.

The explosive progress of computer science and robotics is worrying. Several publications aim at identifying the real stakes of these new technologies.

Le Monde - THE WORLD OF BOOKS | 08.04.2017 with 15:13 | By Céline Henne (Collaborator of the "World of books").

Translated from the French. LE MONDE DES LIVRES | 08.04.2017 à 15h13 | Par Céline Henne.

And if the greatest threat to mankind was not an ecological disaster or the nuclear bomb, but artificial intelligence? The idea that humans may one day live with robots that have become powerfully intelligent or threatened by them has always fascinated the world of science fiction and is still making money in Hollywood, Think of Terminator, by James Cameron (1984), A.I. Artificial Intelligence, by Steven Spielberg (2001) or Ex Machina, by Alex Garland (2015). In recent years, however, it has gone beyond the domain of fiction and has been relayed by scientists and engineers of great renown: astrophysicist Stephen Hawking and billionaire Elon Musk, co-founder of PayPal and SpaceX, have shared with others their concerns about the risks that artificial intelligence causes to our species. For Ray Kurzweil, project manager at Google, the issue is even inevitable: the exponential development of technology signals the programmed end of humanity as we know it, and it is only a matter of decades ...

While a wave of publications on the subject has appeared; the works of Laurence Devillers, Des robots et des hommes. Mythes, fantasmes et réalités (Of Robots and Men, Myths, Fantasies and Realities), and Jean-Gabriel Ganascia, Le Mythe de la singularité (The Myth of singularity). Should we fear artificial intelligence? Both professors of computer science and specialists of the question, are placing themselves subtly against the current of this new trend. For them, it is a question of unraveling the facts of fiction, and of proposing a realistic description of artificial intelligence in order to identify the true stakes. Of course, one can almost feel a little disappointed with this return to reality - some of us prefer to listen trembling to experts predicting the end of Homo sapiens. But let us reassure ourselves: the issues raised by the authors and the technological advances of tomorrow still have enough to provoke many emotions, between wonder and chills.

A prophecy passes for a scientific prediction.

The essay by Jean-Gabriel Ganascia aims at exposing the faults contained in the discourse of the "disaster traders," in particular the announcers of the "singularity", that is to say the moment of rupture that would constitute the emergence of an artificial intelligence capable of improving itself, and which would mark, for the human era, the beginning of the end. He is worried about the great intellectual authority enjoyed by "technoprophets", who contribute to a confusion between genres: what is, at most, a prophecy or a "vast cosmic narrative," passes for a scientific prediction!

The author thus applies himself to showing the limits of their arguments, both empirically and epistemologically, and denounces the confusion between the concepts. These new prophets all predict the passage of "weak" artificial intelligence, the one that is developed at the moment, endowed with autonomy in the technical sense (the ability of a machine to make a decision alone), to a “strong” intelligence, endowed with consciousness and autonomy in the philosophical sense (in which the machine would give itself "the rules and the objectives of its behavior"). However, he said, the prediction of such a qualitative leap cannot be justified in the current development of research.

The greatest danger, according to the two researchers, of these pseudo-scientific myths is that they divert attention from the real stakes present and future. For Laurence Devillers, "the current intelligence of the machines is already very worrying, the announcements shattering on their supremacy must not mask the ethical problems already present". Jean-Gabriel Ganascia thinks that these prophecies lead mostly to disempowering, hiding behind a "unique scenario presented as fatal", which leaves no room for freedom or action for man.

"Social robotics at the service of man".

Laurence Devillers delivers an enlightening insight into how artificial intelligence will overturn our ways of life and social interactions. In the short fiction that opens her essay, no exterminator robot or consciousness, downloaded on a computer, but sympathetic machines that act as aids at home and accompany people on a daily basis. Able to dialogue and simulate emotions, they even manipulate humor. This "social robotics in the service of the human being", on which her research team is currently working, carries many promises: the first versions of these machines have already proved themselves to people with Alzheimer's disease and autistic children.

Far from being optimistic, Laurence Devillers underlines the many ethical problems posed by these technological developments. How to manage the risk of too strong an attachment, or even dependence on these benevolent robots, in vulnerable people? How to ensure the data protection of these machines that will share our privacy? Who is responsible for the actions of a stand-alone robot and its "learner," its designer or its user? It is essential to provide for regulations on their use and regular monitoring, as it is necessary to implement ethical rules in their programs: both technical and legal challenges for tomorrow.

The economic imperatives of the digital giants.

But the main stake remains political. On this question, it is clear that it is the digital giants, the GAFA (Google, Amazon, Facebook, Apple) and a few others who seem to be leading the dance against increasingly powerless states. In the conclusion of his essay, Jean-Gabriel Ganascia accuses these multinationals of practicing a "bewitched charity": under the guise of warning about the dangers of artificial intelligence, they give themselves an image of benevolent enterprise, which can only inflect for the good of all a technological development which, in any case, is inevitable.

According to him, this strategy does not only respond to economic imperatives. Multinationals gradually take over the state's security, biometrics and health functions, which they pretend to assume better and at lower cost. They are creating a new society, the evolution of which they alone can master. So it is not the machines themselves that are to be feared and to be watched, but those who decide their ends and the rules of the game. One can trust man: he knows how to be frightened by himself.