It is possible, by applying artificial intelligence (AI) processing, to access the image of a person lawfully available in the public space and to represent this person naked or in a situation where their dignity is violated. These are unprecedented threats to the basic respect that is due to every human being.
The ability to pervert perfectly legitimate images to humiliate or harass individuals is now within everyone’s reach. The ability to distinguish truth from falsehood is radically undermined. This is where AI has the potential to redefine the rights and freedoms of individuals and communities.
The examples are multiplying and indicate to what extent the neglect of States to regulate the development of these technologies is liberticidal. The show The decryptors of December 2 reported “the growing presence of websites that allow individuals to be stripped naked without their consent, using a simple photo. A turnkey process made possible through AI.”
Radio-Canada journalists tested the functions available on sites that offer to process images in order to transform an ordinary photo into an image of a naked person. They uploaded an AI-created image showing a young woman at the beach wearing a summer dress. Within minutes, the device automatically detected the clothing and removed it, replacing it with the extremely accurate approximation of a naked body. A photo of the type that most people regularly post on the Internet was easily transformed into a nude image.
The most intimate feelings of individuals can also be confronted with manipulations produced by AI tools. In The world on November 30, Nicolas Six reported on the artificial friendship made possible by AI. He reports the words of Thierry Bollon, researcher specializing in social psychology and AI, who assures that “AIs will use all the strategies that we know in business schools to make people buy things that people do not need. Social psychology shows that we are not the decision-maker in an interaction.” He fears that these AIs will be used by promoters of disinformation “to incite terrorism, racism, sexism”.
These are concrete examples illustrating how urgent and justified the calls from several AI experts are to put in place regulations to govern the deployment of these tools.
Review the architecture of rights and freedoms
The capabilities of AI force a review of the architecture of several rights and freedoms. Let’s take the example of image rights, since this site which offers to strip a person appearing in a photo radically calls into question this right. In Quebec and in At various levels of jurisdiction, a person’s image can be disseminated only when that person consents or when the public interest justifies disseminating without their consent.
But with the availability of artificial intelligence tools capable of altering a person’s image, a completely innocuous photo, even one posted with the person’s consent, can become a pornographic icon. If we want to ensure respect for the image rights of individuals, we must now intervene in the circulation of devices and uses of AI.
Now, it’s not realistic to start chasing every image forged by malicious organizations or individuals to try to get them removed from online sites. It is on the circulation of technical devices allowing such manipulations that we must intervene. Otherwise the rights of individuals are just empty incantations.
By default, if states do nothing, AI and other technologies impose their operating rules. AI involves regulations that are imposed on everyone without any real discussion. The real question is whether these rules imposed by technology are compatible with human dignity.
To regulate such powerful technologies, capable of producing images, sounds and texts without regard to their conformity with the truth and respect for fundamental rights, the States must go beyond abstract declarations of principles. We need laws imposing real obligations on companies that implement and use AI-based systems. And when the devices carry risks that we do not know how to assess and manage, we must have the courage to prohibit their circulation.
We cannot imagine that medicines which could have serious consequences on the health of individuals or communities could be legally put into circulation without solid and prior guarantees. Why should it be any different with these devices capable of transforming our vacation images into orgy videos in just a few clicks?
Professor, Pierre Trudel teaches media and information technology law at the University of Montreal.