Footage of a supposed meeting between President Joseph Aoun and Prime Minister Nawaf Salam, generated by artificial intelligence and presented as part of a television news bulletin.
A field correspondent delivering live reports from southern Lebanon, also generated by AI.
Lebanese media outlets are increasingly flooding their coverage with AI-generated videos, to the point where producing news reports and entire pieces of content using the technology has become embedded in newsroom practices.
Is this an embrace of advanced technology, or a reckless rush toward an easy solution whose consequences have not been properly considered?
The widespread push across Lebanese media to rely on AI-generated news content suggests that the institutions adopting these practices are not even asking the most basic questions about journalistic ethics, individual rights, human dignity, and the value of the material they produce as a potential record for the future.
The practice is no longer limited to generating a clip to fill the visual gap created by missing footage in a news bulletin, itself a practice open to criticism. Media outlets are now generating content that depicts not only powerful political figures, but also other segments of society, often portraying them in mocking or negative ways.
Freedom of expression may protect satire. But using AI to fabricate specific situations involving real people demands far greater consideration of what such representations mean and where their limits should lie.
These institutions treat images circulating on social media, and the likenesses of the people in them, as though they have become freely available property. They allow themselves to manipulate them however they imagine or wish, creating whatever scenes, poses and expressions come to mind.
International human rights standards, by contrast, establish safeguards governing the use of individuals’ images, grounded in the rights to privacy and dignity and the protection of reputation and personal identity. These safeguards are intended to preserve people’s rights and human dignity. Yet some media institutions are creating images that depict individuals in degrading situations, treating their humiliation as a source of entertainment for both those producing the content and those consuming it. This is particularly evident in satirical AI-generated videos.
For example, LBCI generated a video depicting well-known content creator George Dib, known as “Dr. Food,” behind prison bars over drug-related charges. His ex-wife, known as Shorouk, appears in front of him carrying sweets and a bundle of bread as he begs her to get him out of prison. She waves goodbye mockingly and says, “Bye, Dr. Sauce,” to which he replies, “I wish I’d sold bedsheets instead of pillows.”
Beyond depicting the two individuals with expressions they never actually made, the video humiliates one of them while fabricating words he never said.
“He Should Be Portrayed in the Ugliest Possible Way”
On social media, individuals generate even more extreme scenes and insults targeting female journalists whose political views they disagree with. These shocking, misogynistic depictions reveal a mindset willing to resort to virtually any form of attack, one that becomes even more vicious when the person being targeted is a woman.
One of the key problems raised by a video such as the “Dr. Food” clip is that an amateur did not create it. An established media institution with a long history in Lebanon published it.
Hassan al-Khatib, who produces AI-generated segments, quickly agreed to answer Daraj’s questions about the controversy surrounding the content he creates for LBCI, the ethical issues it raises, and the boundaries he sets when depicting real people.
Al-Khatib says the applications he uses impose restrictions that “do not allow you to go too far.” Asked about manipulating the images of people who are not in positions of power, such as George Dib and his ex-wife, and the ethical questions this raises, he says: “I am not doing anything unethical. I am exercising freedom of expression.”
“They air their dirty laundry on social media and share every detail of their personal lives, and Lebanese people attack them,” al-Khatib continues. “Dr. Food is a drug dealer who harms people. He and his ex-wife used to threaten anyone who criticized them. What is happening today is karma. As far as I’m concerned, he should be portrayed in the ugliest possible way.”
Al-Khatib says his content is meant for social media and is “close to the Lebanese street,” adding that the expression “ya sauce” is commonly used on these platforms.
The content creator frames his criticism within the broader context of attacks against LBCI after an earlier incident in which he produced an AI-generated video mocking Hezbollah Secretary-General Naim Qassem, sparking widespread controversy. Al-Khatib says the broadcaster’s management reviews his work and that he abides by its decisions. He argues that the Ministry of Information is responsible for regulating such practices.
He also says that he and members of his family have received death threats, while those making the threats have faced no accountability, forcing him to take precautions.
This is precisely where the responsibility of Lebanon’s legislative and executive authorities comes into focus. They have consistently lagged in passing legislation that keeps pace with technological change. Until the recent adoption of the new media law, online news platforms had operated for more than two decades without a proper regulatory framework.
Deepfakes Threaten the Visual Archive
Returning to AI-generated video, Lebanese media institutions are increasingly using artificial intelligence to produce news reports, some of which could now be described as deepfakes.
Al Jadeed, for example, generated footage depicting the arrest of Dr. Food for a news report about the case, even though the report could easily have been produced without such a scene. The footage could arguably be considered a fabrication, even though it was labeled as AI-generated. Here again, an ethical problem emerges: a real person is depicted being arrested as the report’s producer imagines the scene, within a news bulletin that is supposed to document reality, not invent it.
Several months earlier, the same broadcaster generated scenes depicting singer Fadel Shaker appearing before a judge at the Military Court. LBCI, for its part, generated images showing Bashar al-Assad behind bars for a news report, even though he has never been imprisoned.
Among other examples, Al Jadeed also aired a report on smuggling whose visuals were produced using artificial intelligence.
This practice raises a fundamental question about whether news organizations are disregarding the importance of images as historical documents, abandoning them too readily in favor of fabricated visual representations and invented scenes. It also reflects an increasingly casual reliance on artificial intelligence, a trend many experts warn could steadily contaminate the visual archive and distort the historical record for future generations.
These fabricated scenes fulfill the desires of those who create them, depicting not what actually happened, but perhaps what their creators wish had happened. The excessive use of AI-generated imagery in news coverage introduces artificial representations of real historical events into the visual archive, effectively falsifying history.
It could even be argued that such imagery is increasingly being used to satisfy a voyeuristic impulse: the desire to see people depicted at their most vulnerable.
What If Lies Had Convincing Images?
Fabrication in news bulletins is likely to become even more dangerous as AI tools grow increasingly sophisticated, capable of reproducing ever finer details while becoming more accessible.
Media owners in Lebanon and around the world have extensive networks of interests, and these tools could make it easier for them to disseminate inaccurate information that serves those interests. Many already do so today, but the possibility of exposing falsehoods still constrains them. As audiences become more inclined to believe what they see, and as the technology makes it increasingly difficult to distinguish fabricated reports from authentic ones, news could shift from documenting reality to constructing a scenario, generating the images to support it, and ultimately convincing audiences that it actually happened.
Reporters Without Borders (RSF), together with 16 partner organizations, launched the Paris Charter on AI and Journalism. Among the principles it identifies as essential to using artificial intelligence in journalism are accuracy, fairness, harm prevention, non-discrimination, accountability, and respect for privacy. The charter also holds media organizations accountable for all content they publish, including content generated using artificial intelligence.
It calls on news organizations to prioritize authentic images and recordings when depicting real events and, in particular, to refrain from using synthetic content that mimics real-world recordings or realistically depicts real people.
In Europe, the European Union has faced tensions with major technology companies over its efforts to introduce safeguards protecting users. It has established a legislative framework governing artificial intelligence across the EU, with some of its provisions entering into force at the beginning of last August. Under the legislation, labeling content as artificially generated does not exempt its publisher from complying with other laws governing the use of images.
Ultimately, this is not a debate between modern media and those resistant to technological change. Artificial intelligence has already become part of the journalistic toolkit, and its role will undoubtedly continue to grow. The question is where the line lies between using AI as a tool that helps journalists understand and present reality, and using it to fabricate a reality that never existed.
In news, especially, an image is not merely a visual element used to fill a screen. It is a document, a testimony, and part of our collective memory.
When realistic images are created to depict an arrest that was never filmed, a trial for which no footage exists, or a political meeting from which no images were released, they can no longer be regarded as mere visual illustrations, even when labeled as AI-generated.
What we are dealing with, in such cases, is an artificial reconstruction of reality, one that will enter the visual archive and risks being recirculated in the future as a record of something that actually happened.
In a country like Lebanon, where new technologies often outpace legislation and regulation, media organizations cannot wait for laws and regulatory frameworks before setting boundaries for themselves. Responsibility begins with editorial leadership and media management. News organizations must prioritize authentic images and refrain from fabricating scenes of real events that were never filmed, except in exceptional cases where clear editorial necessity exists.
What is at stake is not only the credibility of a news story or an image broadcast today, but also the kind of archive and historical record journalism will leave behind tomorrow. If the media grows accustomed to inventing images of reality whenever the camera is absent, the problem will not simply be that artificial intelligence has become better at imitating reality. It will be that journalism itself has abandoned its responsibility to protect it.
