[La chronique de Pierre Trudel] War and its technological overkill

For a long time, the truth has been among the first victims of war. What is unprecedented is the availability of technologies that can generate and disseminate words and images that have nothing to do with verifiable reality on the ground. The making of deep fakes (deepfakes) is now within the reach of many people. It is a weapon in the propaganda battles inherent in contemporary wars. To protect yourself from disinformation, you have to give yourself the means to decide between the true and the false.

Hyperfakes are in particular computer-generated images produced by mobilizing technologies based on artificial intelligence. Professor and doctor in computer science Solange Ghernaouti explains, in her blog at Time, that it is “true-false” information, created to manipulate public opinion. The illusion of truth is total. The user cannot differentiate the true from the false with the naked eye. The hyper-faked document is made by computer coding, without being based on real and verifiable facts. Reality does not need to exist. The only thing that counts is the existence of digital content distributed, which is intended to be seen, listened to, shared and above all liked!

Legitimate and malicious falsehoods

These tools, which work thanks to deep learning and artificial intelligence techniques, are capable of the best and the worst. They can be used both by individuals seeking in good faith to present themselves in a better light, and by fraudsters and stalkers disguising themselves under false identities. These technologies are mobilized to feed political and war propaganda. They make possible the mass distribution of deliberately falsified documents. This is a serious threat to the integrity of democratic deliberations.

Devices for making fake images and videos can serve legitimate purposes. Fiction films and video games make extensive use of these technical capabilities. Banning them outright is not a realistic avenue. We must rather sophisticate our means of distinguishing the true from the false.

To avoid placing state authorities in a position to determine what is “truth”, our laws refrain in principle from deciding between what is true and what is false. There are, however, important exceptions. Misleading advertising laws penalize the dissemination of claims that cannot be supported by scientific or technical knowledge. False or misleading statements that damage the reputation of individuals or those that are part of racist or xenophobic discourse can be punished under the laws. The same goes for fraudulent practices. But apart from these explicitly harmful activities, how to decide between the true and the false?

The political debate unfolds on the basis of perceptions of reality. These perceptions are necessarily differentiated in particular according to the experiences of each other, their interests and their visions of the world. But there is a fine line between the questionable and the fake. Presenting as authentic a video of a political figure produced using artificial intelligence technologies by making him declare statements contrary to those he has made on many occasions clearly constitutes falsification.

The challenge for laws and justice systems in this world where information travels at high speed is to differentiate legitimate falsehoods from those that are not. It is important to take the trouble to distinguish between falsifications motivated by an intention to deceive and those which proceed from a humorous or otherwise playful approach. In a democracy, we must rightly beware of punishing remarks based on simply questionable foundations. For a law prohibiting falsehood to be considered a reasonable limit on freedom of expression, there must be a convincing demonstration of the dangers that must be averted. Prohibitions must be proportionate to the evils to be combated.

What to do ? A report on the challenges of the synthetic society written by researchers from the University of Tilburg, in the Netherlands, recommends regulating the use of processes based on artificial intelligence. The researchers note that it is necessary to set up legislations regulating or prohibiting the production, the distribution or the possession of devices intended to produce deepfakes. We also put forward more stringent prior verification requirements for online platforms, such as Twitter or TikTok. The importance of strengthening laws against disinformation practices in elections and counterfeit evidence in court is stressed. To limit the propensity to forge and falsify facts and gestures attributed to well-known personalities, we call for the strengthening of protections for the privacy and image of individuals.

But even if we have the strictest laws against deception, it is a free, pluralistic media capable of verifying information that constitutes the best bulwark of truth. In the face of information fueled by deep fakes, independent media and their journalists who work on the ground according to rigorous standards are more essential than ever.

To see in video


source site-40

Latest