# AI on official accounts: viral play, threat or distortion? Author: Martín Sosa Dirié Original publication: 2025-03-26 Spanish original: https://estrategiabyaleph.substack.com/p/estrategia-78-de-trump-a-macron-como English URL: https://elcontemplador.github.io/estrategia-english/essays/078/ Status: Published translation This is a translation of the original Spanish essay published on 26 March 2025. Its claims, examples and forecasts retain that historical context. English publication: 2026-09-29 *This week, our main article once again features a contribution from [Martín Sosa Dirié](https://www.linkedin.com/in/martinpsosa/), a journalist, communications consultant and graduate of the Master's in Image Advising and Political Consultancy directed by Jorge Santiago. He also wrote [the interesting lead article in issue 67 of estrategIA](https://estrategiabyaleph.substack.com/p/estrategia-67-predecir-para-influir).* Much of the debate about artificial intelligence and political communication has centred on its potential risks, especially the spread of disinformation through artificially generated images, videos and voices. But to what extent have these warnings materialised? Has AI established itself as an effective tool for manipulating public discourse, or has its impact taken unexpected directions, creating new forms of risk and distortion? I invite you to join me in reviewing some recent cases in which the **official accounts** of presidents and political parties have used AI in their communications. ## 1. AI: the bizarre and its impact On 26 February, Trump once again shook the international agenda with a social media post. This time, it was not an affront, a threat or a sarcastic criticism of an opposing politician, as we have grown accustomed to, but the arrival of the “Trump Gaza” reel. [realdonaldtrump](https://instagram.com/realdonaldtrump) [![Retained still from the AI-generated Trump Gaza reel, depicting a monumental golden statue in an imagined resort city.](https://elcontemplador.github.io/estrategia-english/assets/images/9952f0ae-8528-4650-8db1-f64b94c0c26d_640x1138.jpeg)](https://substackcdn.com/image/fetch/$s_!UHWB!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9952f0ae-8528-4650-8db1-f64b94c0c26d_640x1138.jpeg) *A post shared by [@realdonaldtrump](https://instagram.com/realdonaldtrump). The retained still shows an AI-generated golden statue in the reel's imagined resort setting; it is not a photograph of a real development.* The video lasts **33 seconds** and begins with images of a ruined city in which children are pursued with rifles. The caption **“Gaza 2025/What's next”** locates the images in that area of the Middle East, before switching to a sequence of scenes of Gaza transformed into a kind of holiday paradise, with clear-water beaches, white sand, gold statues of Trump, five-star resorts and even a Trump Tower. The short features **Donald Trump**, **Elon Musk** and **Benjamin Netanyahu**, with a soundtrack whose lyrics say: > “Donald is coming to set you free, bringing life for all to see. No more tunnels, no more fear, Trump Gaza is finally here. Trump Gaza shining bright, a golden future, a whole new light. Feast and dance, the deal is done, Trump Gaza number one.” *Translation note: these lyrics are translated from the Spanish quotation reproduced in the original article; they have not been verified against the original audio.* The piece exceeded **60 million views** on Instagram alone, **twelve times the average** his regular content had been receiving there. ## 2. AI as an extension of personal branding and a reinforcement of identity Another interesting case is Javier Milei, probably **the serving president who makes the greatest use of AI tools**. Since his 2021 election campaign, when he built a **strong personal brand** associated with [the figure of the lion](https://youtu.be/OTEP6eZmzx0?si=nWW6beKwAKq4Y5fH)—projecting strength, courage, boldness and leadership—Milei has continually [reinforced that association](https://www.instagram.com/p/C20Hp9BAP18/?utm_source=ig_embed&utm_campaign=loading) through AI. He uses these tools frequently to depict a wide variety of political events, whether [defending a government measure](https://x.com/JMilei/status/1798216123240132715), marking [a bilateral meeting](https://www.instagram.com/p/DCBqGQuxCqP/) with a foreign president, or portraying his government's heroism and bravery through playful references to popular-culture figures: Terminator, Superman, Rocky, Neo from *The Matrix*, and so on. [![Two screenshots from Javier Milei's account featuring synthetic Terminator and superhero imagery; post captions, image slogans and visible comments are translated below.](https://elcontemplador.github.io/estrategia-english/assets/images/ca0bc26a-3c0f-4e57-9c52-297d662e3ea8_1456x383.jpeg)](https://substackcdn.com/image/fetch/$s_!jk90!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fca0bc26a-3c0f-4e57-9c52-297d662e3ea8_1456x383.jpeg) *Translation of the substantive text in this montage of posts from Javier Milei's account:* - **Terminator image:** “Belliboni detected. Trade unionist detected. Governor detected. Member of parliament detected. Caste in sight, baby.” “Caste” refers here to the political establishment in Milei's rhetoric; the last line plays on a Terminator catchphrase. The artwork bears the signature “Nik”. - **Caption beside it:** “The caste is trembling. Freedom advances. Long live freedom, damn it.” - **Visible comments beneath that caption:** “How you lulled us with the lies you promised, eh!!! In the end I was the caste, you petty thief”; “When your term ends in 2027, I'm going to post ‘SE TERMINATOR’”—a pun on *se terminó*, “it's over”; “Hey, hasn't this meme aged badly? In the end the caste was the people and the goblin came to sell EVERYTHING THAT ‘DOES NOT BELONG TO HIM’. In other words, another thief, only a white-gloved one. Still a thief”; and “Now NIK is lending himself to this nonsense.” - **Caption beside the superhero image:** “Impossible not to adore my followers. And now they are bringing in the magic of AI as well. Long live freedom, damn it.” - **Visible comments beneath it:** “They went to vote three times and this seemed the best choice”; “Urgent therapy”; and “I want to defend you, but you're not helping.” *These translations document the posts and reactions visible in the original screenshot; the statements are those of their respective posters.* He has also used AI to reinforce strategic points in his narrative: the defence of [freedom](https://www.instagram.com/p/DEzxGiYxfB2/?utm_source=ig_web_copy_link&igsh=MzRlODBiNWFlZA==); [“awakening lions”](https://www.instagram.com/p/CnPyCwqvprw/?utm_source=ig_web_copy_link&igsh=MzRlODBiNWFlZA==); and the fight [against communism](https://www.instagram.com/p/C2OnCSMoRZR/?utm_source=ig_web_copy_link&igsh=MzRlODBiNWFlZA==) and [against the caste](https://www.instagram.com/p/CyTeaC3Ae2K/?utm_source=ig_web_copy_link&igsh=MzRlODBiNWFlZA==), among others. [![Synthetic lion imagery in Javier Milei posts: a lion guarding cash and a lion with a mug labelled Leftists' tears; the visible slogans are translated below.](https://elcontemplador.github.io/estrategia-english/assets/images/398c7cf0-b19f-40ea-a4a7-11cb2f473887_1456x921.jpeg)](https://substackcdn.com/image/fetch/$s_!HF0N!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F398c7cf0-b19f-40ea-a4a7-11cb2f473887_1456x921.jpeg) *The left post, from @JMilei, reads “Defending the cash box…” and depicts a lion in ceremonial dress with a sword, banknotes, rats and the Argentine flag. The right-hand image shows a lion drinking from a mug labelled “Leftists' tears”. Both are synthetic illustrations used in the original article as examples of political messaging.* It is also worth recalling the **2023 presidential campaign**, when **Milei depicted** one of its **key moments**, his alliance with Patricia Bullrich for the run-off against Sergio Massa, through an **AI-generated image**. [![Screenshot of a 25 October 2023 Javier Milei post showing an AI-generated lion and duck embracing, representing the alliance described in the article.](https://elcontemplador.github.io/estrategia-english/assets/images/6a28a847-1c73-4667-8eac-aaac9f60a217_540x638.jpeg)](https://substackcdn.com/image/fetch/$s_!1Isu!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6a28a847-1c73-4667-8eac-aaac9f60a217_540x638.jpeg) *The original post from @JMilei is dated 25 October 2023 at 13:16. It shows the synthetic illustration discussed below.* The image, showing a duck representing Bullrich and a lion representing Milei embracing, illustrated the political alliance in a novel way. ## 3. AI as a tool for political denunciation and promotion Another recent case with considerable impact was **launched by Spain's Partido Popular (PP)** at the beginning of the month to denounce alleged corruption within the PSOE and Pedro Sánchez's government. Entitled **“The Island of Corruption”**, the campaign used an AI montage to **recreate** the opening of the popular television programme **“Temptation Island”**. In the video, political figures associated with the PSOE, including Pedro Sánchez, his wife Begoña Gómez and others, are introduced as “contestants” on an island where “corruption, lust, ambition and lies” reign. The video **exceeded two million views** on the PP's **X** account in its first 24 hours, **multiplying by 75—yes, 75!—the views of some** of the same account's most-watched videos that week. Something similar happened on **Instagram**, where it accumulated **more than 10,000 likes and 350,000 views** in just one day, between eight and ten times the average. The PP later deleted it following an oversight that prompted [a complaint from the Dominican Republic's embassy](https://elpais.com/espana/2025-03-06/el-pp-borra-el-tuit-sobre-la-isla-de-las-tentaciones-tras-la-queja-de-republica-dominicana.html). **Macron**, too, recently experimented with AI by publishing a video full of **deepfakes of himself** to **promote the AI Action Summit** held in Paris in February. [emmanuelmacron](https://instagram.com/emmanuelmacron) [![Still retained from the Macron deepfake video used to promote the February 2025 AI Action Summit.](https://elcontemplador.github.io/estrategia-english/assets/images/949b1422-194e-4c1d-8196-90461e867b8a_640x1136.jpeg)](https://substackcdn.com/image/fetch/$s_!1_-G!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F949b1422-194e-4c1d-8196-90461e867b8a_640x1136.jpeg) *A post shared by [@emmanuelmacron](https://instagram.com/emmanuelmacron). The retained still is from the deepfake video discussed in the article; it is not presented as an authentic scene.* It reached **8.4 million views**, **eight times** the reach of his most-watched reels. ## 4. Beyond volume As the cases examined show, AI—at least so far—has not been used in political communication with the aim of deliberately deceiving or spreading disinformation. Not, at least, from official accounts, which are what we have explored here. **The applications we have seen tend to focus on creating exaggerated, surreal and fantastical scenes that do not seek to pass themselves off as real. Instead, they aim to exploit the humorous, satirical or memetic avenues of digital culture.** This content seeks to amplify its impact through virality—and succeeds—by taking situations to the limit or even beyond, creating material that borders on the bizarre, the disruptive and the extravagant. Yet evaluating these pieces solely from a quantitative perspective, such as their volume of views or plays, is not enough. Every piece of political communication serves **strategic objectives**, both **communicational** and **political**. Although AI creations clearly achieve greater reach in terms of views and engagement, the **key question** is whether this kind of content contributes to political discourse or, instead, **degrades and trivialises it**. It is therefore essential not to allow qualitative analysis to be eclipsed by the dominance of volume and virality. Indeed, a broader, more systematic study combining quantitative data—views, plays, shares, engagement and so on—with in-depth qualitative analysis would be highly valuable. It would allow a more rigorous assessment of AI-created pieces in political communication, not only in terms of reach, but also in relation to the quality and meaning of the political message they convey. [Martín Sosa Dirié](https://www.linkedin.com/in/martinpsosa/) *Journalist and communications consultant* Author of [@conversacioncontinua](https://conversacioncontinua.substack.com/).