Roger-Luc Chayer (Image : AI / Gay Globe)
Nothing Is Right Anymore on the Social Media Planet
Nothing is right anymore on the social media planet, as a recent investigation revealed that more than 20% of content broadcast on YouTube, a property of Google, would have been generated by artificial intelligence and corresponds to no tangible reality.
The entire social media ecosystem is now affected by this cancer, turning the information ecosystem into veritable quicksands. This drift calls into question everything humanity has built so far in terms of knowledge and truth, notably managing to convince millions of people that the Earth is flat, that man never walked on the Moon, that HIV does not exist, or that climate change is an invention by subsidized “woke” activists spreading false information.
Threads, Champion of Disinformation
Champion of disinformation belonging to Meta (Facebook, Instagram), the newest member of the family publishes, by the millions — about 100 million messages per day — absolutely anything.
For 30 days, I have precisely observed the nature and origin of messages shared on this social network and, by my estimate, more than 50% of what is published is false. As for the other half, corroboration is so complex that, until now, I have only managed to confirm about five posts out of the hundreds analyzed.
According to ChatGPT, “the high rate of disinformation and fake news on Threads is primarily explained by the platform’s very architecture and the strategic choices made by Meta at its launch. Threads was designed as a fast conversation space, highly interconnected with Instagram, promoting virality, instant reaction, and the circulation of short messages, often posted without prior verification. This kind of environment favors speed and emotion at the expense of contextualization and informational rigor, two essential elements for content reliability.
Added to this is the initial absence of robust proactive moderation and fact-checking mechanisms. This relative permissiveness created fertile ground for the spread of misleading, approximate, or deliberately false content, especially in political, scientific, and societal fields.
The recommendation algorithm also plays a central role. As on other social platforms, it tends to amplify content that provokes strong reactions, whether outrage, fear, or confirmation of pre-existing beliefs. Disinformation is particularly effective at capturing attention, simplifying complex issues, and offering divisive narratives. In this context, factually false but emotionally powerful content benefits from disproportionate visibility.”
YouTube, a Major Vector of False Information
Several fact-checking experts recently revealed that the video platform YouTube, owned by Google, contained an alarming amount of false information and videos created by artificial intelligence, to the point where they estimated that up to 30% of content broadcast was false or misleading.
The British newspaper The Guardian already published in 2022 a call for vigilance addressed to authorities to better regulate what is broadcast on YouTube. In an open letter signed by 80 international organizations, it was mentioned: “YouTube is a major vector of disinformation worldwide, according to a coalition of more than 80 fact-checking organizations like Full Fact (UK) and the Washington Post Fact Checker. In a letter addressed to CEO Susan Wojcicki, these groups denounce the spread on the platform of false content, notably related to Covid-19 and the theory of American election fraud. They believe YouTube enables malicious actors to manipulate users, and that current measures are insufficient.
The coalition calls on YouTube to fund independent research on disinformation, add links to refutations in misleading videos, prevent its algorithms from promoting repeat offenders, and strengthen the fight against disinformation in non-English content. They particularly highlight the problem in the ‘Global South’ (Latin America, Asia, Africa), where disinformation is widespread.
YouTube claims to have invested in policies to limit the spread of misleading content, notably by removing videos violating its rules, such as those on Covid vaccines.” However, in 2025, the spread of artificially created videos exploded.
Facebook, a Giant of False Promises
Facebook, belonging to Meta like Threads and Instagram, is the largest social network in the world. According to the latest statistics available in 2025:
- More than 350 million posts are created daily on Facebook worldwide. This includes statuses, photos, videos, links, and other types of content shared by users.
- In terms of interactions (likes, comments, shares), the daily total often exceeds 1.5 billion. These interactions reflect user engagement with published content.
- Facebook has approximately 2.9 billion monthly active users, with a large majority logging in daily, contributing to these very high volumes.
I don’t know if, like me, you have noticed recently, but this social network has become the master of spreading false political and medical information. Thousands of accounts continuously share news on all sorts of medical conditions, provoking millions of reactions. Everything goes: curing cancer, hair regrowth, eradicating HIV, natural tooth regrowth. The worst is reading people’s reactions rejoicing and endlessly sharing these falsehoods.
Facebook claims to have implemented mechanisms and algorithms to filter accounts spreading false information, but I strongly doubt this claim. My daily news feed, at the moment I write this article, contains a news story about Céline Dion’s new tour (false), an announcement from an American lab claiming to reverse all lung damage caused by smoking (false), a rumor that Céline Dion threatens Sony with legal action (false), or even the idea that orcas carry dead salmon on their heads like hats (false). And this is in less than one minute of scrolling.
Grok, a Global Scandal
The newest social network belonging to Elon Musk has been involved for several weeks in a large international investigation as the artificial intelligence tool linked to the X network is accused of creating and spreading artificially made videos showing child pornography.
According to The Guardian, the mother of one of Elon Musk’s sons said she felt “horrified and violated” after fans of the billionaire used his AI tool, Grok, to create fake sexualized images of her by manipulating real photos.
Screenwriter and political strategist Ashley St Clair, who distanced herself from Musk after their child was born in 2024, told The Guardian that supporters of the X owner used this tool to create a form of revenge porn, going so far as to undress a photo of her when she was a child.
Grok has been sharply criticized by legislators and regulators worldwide after it was revealed that it had been used to virtually undress images of women and children, showing them in compromising sexualized positions. These massive sexual abuses involve X users asking Grok to manipulate photos of fully clothed women to depict them in bikinis, kneeling, and covered with what looks like semen.
Solutions to Get the Web Back on Track
China and France have tried to implement mechanisms to manage, detect, and punish those who spread false information, and it works particularly well. The example of the First Lady, wife of Emmanuel Macron, is telling. Subject to rumors about her gender for years on the Internet, Madame Macron obtained the arrest of eight people according to the French criminal code, and these eight individuals received prison sentences.
The catch is that there is no universal rule regarding web usage. The Internet was originally made public to allow broader access to a network until then reserved for military and academic uses. The idea was to promote information sharing, collaboration between researchers and institutions, and stimulate technological innovation by opening this network to a larger number of users. This transition to public diffusion also aimed to democratize access to information and create a universal communication space capable of connecting diverse communities worldwide, thus contributing to economic, scientific, and social development.
But we are no longer at all there. The web has become a tool for spreading falsehoods and hate.
One of the means some groups of academics are trying to develop is to manage the web exactly like the road. To drive a road vehicle, regardless of the country, you need a driver’s license. The same could be required to navigate the web.
If a license to circulate on the web were to be implemented internationally, its operation could be inspired by physical driver’s licenses, with several key steps. First, there would need to be standardized training worldwide to raise awareness among internet users of good digital practices, such as security, privacy protection, fighting disinformation, and respecting laws in force. After this training, an exam would probably be necessary to validate the knowledge acquired, allowing the issuance of a unique digital license recognized by different countries.
This license would then serve as proof that the person is capable of navigating responsibly and securely on the internet. To be controlled, it could be linked to a verified digital identity, raising important issues around confidentiality and individual freedoms. This system could also include license levels, adapted to uses — for example, a basic license for browsing and an advanced license for managing content, making online transactions, or administering platforms.
To ensure its effectiveness, this license should be integrated into global digital infrastructures, with mechanisms to suspend or revoke authorization in case of harmful or illegal behavior, while ensuring transparent governance respectful of fundamental rights. This model would represent an ambitious attempt at global regulation of a space as fluid and complex as the web.
This concept would be relatively simple to manage, like IP addresses. To implement such technology, states will need to be convinced that the internet is about to implode under the weight of disinformation and that the tool, as it stands, is no longer credible. For now, this remains to be discussed.
ADVERTISING
