© is463940 / Pixabay
Философия и съвременност
ANTHROPOLOGIZATION OF THE ARTIFICIAL INTELLIGENCE IN THE HYPERREAL SPACE
https://doi.org/10.53656/phil2026-02-06
Резюме. The article explores the trend towards anthropologization of artificial intelligence. It analyzes this trend from the perspective of Baudrillard’s concept of hyperreality, emphasizing the replacement of the real with the signs of the real and, consequently, the trivialization of the distinction between lie and truth, between real and imaginary. In such a context, various modes of AI anthropologisation are examined – specific concepts describing conditions of AI, images of models with their own biographies on social networks, virtual partners who evoke feelings in their creators and prompt them to certain actions. The study puts forward the hypothesis that the process of AI anthropologisation runs parallel to that of human machinization. Examples of similar machinization are given, focusing mainly on the weakening of some human cognitive abilities and the development of others based on the machine model, in which, however, algorithms perform far better. The question of the erosion of human relationships in this situation is raised, and consequently – of the beginnings of totalitarian thinking, of destructive and authoritarian behavior, which isolation and the feeling of insecurity provoke.
Ключови думи: artificial intelligence; anthropologisation; machinization; hyperreality
I. Simulacrum and Hyperreality in Baudrillard Conception
Jean Baudrillard’s famous book “Simulacra and Simulation” begins with a reference to Borges’story “On Exactitude in Science,” in which the cartographers of the Empire, attempting to represent all the details of its territory, create a map as large as the Empire itself and, in fact, coinciding with it. With the decline of the Empire, the map also disintegrates: only a few pieces in the deserts can be deciphered. The charm of this allegory, according to Baudrillard, is that it recalls an already annulled symbolic order, in which signs aimed to reflect space, substance, reality. Today’s simulation is not a copy of an original object, but “the generation by models of a real without origin or reality: the hyperreal” (Baudrillard, 2019, рp. 7 – 8). In this sense, the territory no longer precedes the map, but it is also impossible to say the opposite – that the map precedes the territory. In fact, what has happened is the disappearance of the distinction between the two: between the map and space, between concept and reality, between essence and phenomenon.
The simulacrum, Baudrillard argues, is the final phase in the development of the sign, which does not aim to designate reality, nor even to distort it into a false representation, but to entirely replace it. In the age of simulation, referent objects are abolished, but at the same time – artificially resurrected in systems of signs. In this way, a process of deterring the real is set in motion, which is replaced by “its operational double” – the signs of the real, so that it is no longer possible to distinguish between simulation and reality. They are indistinguishable in hyperreality, which renders meaningless the distinction between lie and truth, between real and illusory.
The seemingly abstract philosophical-semiotic theory of the simulacrum actually has diverse political, social, and cultural dimensions, which Baudrillard analyzes. In the hyperreal space of the modern world, “Apocalypse Now” cannot be distinguished from the Vietnam War, the Romanian revolution and the trial of Ceaușescu represent television speculation, the Gulf War was made by and for CNN, the Lascaux caves are closed, but an exact replica has been built next to them for visitors, Native Americans in America are demographically increased in an attempt to restore a culture whose specificity is precisely in the limitation of the group and the prevention of its free expansion. Disneyland is the model of America, but Disneyland is also its principle – in Baudrillard’s concept, such spaces that self-admit to being simulations aim to hide that America itself, and perhaps the entire world, are hyperreal, that “outside the boundaries of the artificial perimeter there is as much reality as within them.” (Baudrillard, 2019, р. 25).
The images of media and advertising are infiltrated into the real, so that it is impossible to identify cause and effect: information coincides with the event, the image generates events. Events, according to Baudrillard’s terminology, “go on strike,” their place is taken by occurrences determined by media, cinema, and advertising. As in the situation with the symptoms of the imaginary madman, which transform the simulation of madness into madness, or similar to the false robbery that potentially leads to very real consequences, “artificial signs” intertwine with “real elements” (Baudrillard, 2019, р. 34). Of course, incidents continue to occur in modernity, but their authenticity cannot be achieved. In the homogeneous solution of images and recurrences of the real, reality cannot be distilled from the flow of signs. The hyperreal lies precisely in this: images are more real than the real, or rather, the real acquires its reality through the image.
IІ. Interlude on the Realized Dystopia
In fact, the idea of a reality lost or losing itself in simulation, quite independently of Baudrillard’s thematizations, but with an unexpectedly analogous sound, found its place in literature as early as the mid-last century. In 1963, the Italian science fiction writer Lino Aldani published the short story “Good Night, Sofia,” better known as “Onirofilm” (Aldani, 1980). The story describes a near future where people live alone in cramped rooms, have no sexual relations, but are born in vitro under the strict genetic control of the Bank of Life, work for a few hours, and spend the rest of their time dreaming vivid dreams with the help of onirofilms. An onirofilm is a dream experienced as reality; a film in which the viewer becomes a participant in incredible adventures and sexual scenes with alluring women. The main character in the story, Sofia Barlow, is a popular actress who begins to take an interest in the time before onirofilms, to doubt the existing system and way of life, and to seek recognition of her attractiveness in real life, not just in her film portrayals. On a flight to San Francisco, the plane sent to take her to a shoot crashes. The crash, as well as the place where the pilot Mirko Gligorich lands the plane, literally resemble scenes from one of her films. The two must spend the night in a cave: he isolates himself with his personal collection of Sofia’s onirofilms, while she, contrary to all norms, tries to tempt him to choose her over her image in the films. Finally, Mirko throws himself at Sofia and the lights of the amplex (the device for connecting onirofilms) light up: Mirko’s choice turns out to be a scene from a unique, personalized onirofilm, given to her by producer Bradley. When she calls to thank him, Bradley summarizes the spirit of the new oniro-ideology: “The last obstacle to overcome is vanity; you imagined that a man could prefer you to a dream, you fell into the most dangerous heresy, but we noticed it in time and hastened to take measures... Don’t forget, nothing can replace a dream. And only in a dream can you imagine the opposite” (Aldani, 1980, р. 39).
While the theme of simulation appeared sporadically in individual science fiction works until the 1980s, the emergence and spread of the Internet, the development of electronics and communication technologies, and discoveries in the fields of genetic engineering, plastic surgery, and transplants in recent decades have transformed it into a common motif in both literature and, especially, cinema. More than 30 years after “Onirofilm,” Kathryn Bigelow’s film “Strange Days” (1995) unfolds a similar plot. Its characters use a special device called SQUID (Superconducting Quantum Interference Device), which recreates their happy memories, gradually replacing reality. With excessive use of SQUID, the user’s brain cannot withstand it, and they fall into a state close to clinical death. This was followed by “The Truman Show” (1998), “eXistenZ” (1999), “The Matrix” (1999), “The Thirteenth Floor” (1999), as well as numerous successors in the new millennium, enriched with many new plotlines and effects, some of the most famous of which are: “Avatar,” “Inception,” “S1m0ne.”
In most of the aforementioned cinematic works, the simulation is unmasked. Despite the characters’ addiction to it or their direct implantation within it, in one way or another, there is a return to (or restoration of) an authentic reality. Baudrillard’s hypothesis, however, is far more radical. “Onirofilm” tells of a gray world where only the dream-films are in color, providing a perfect illusion of exciting experiences. Baudrillard claims that we already inhabit this world – the onirofilm has modified reality into the hyperreal.
III. Anthropolоgization versus Machinization
This classic and somewhat overused theory was recalled with the conviction that if, when it was created about half a century ago, it represented more of a provocative conceptual scheme that hyperbolized some trends in the development of the modern world to an extreme, then progress in the field of AI seems to give it unexpected completeness and credibility (Tihomirova, 2025, p. 22). It is enough to mention Deep Fakes to see the literalization of Baudrillard’s metaphor of hyperreality: “When there is a fake image of, for example, a famous politician, which is not just a static image, but a dynamic image, there is a video that one cannot distinguish from a real video shot with an ordinary camera. This is already something very different from what we have seen so far as attempts to replace digital identity. The voice of a person, the image of a person, the movement of a person, the words of a person. This is Deep Fake” (Grozdanov, 2025).
However, this text will not focus on Deep Fakes, but rather on other phenomena that will be defined here as the anthropologization of AI and which explicate the thesis of a substituted reality. In 2023, in Barcelona, Ruben Cruz, the founder of the advertising company The Clueless, went through a difficult period. By his own assessment, most of the agency’s problems were not related to the design of the advertisements it offered, but to the irresponsible attitude of models and influencers towards their commitments. And so Cruz created Aitana – an AI influencer with pink hair and a perfect body. In the first month, Aitana acquired a biography, as her creator realized that not only perfect appearance but also stories sell. Accordingly, on the agency’s website, she is described as a fitness enthusiast, determined and with a complex character, sociable and caring. Literally following the plot of the film “S1m0ne,” Aitana gathered over 120,000 followers on Instagram in a few months, as well as invitations for conversations and meetings from celebrities who were unaware of her virtual nature. The model advertises lingerie and also became the face of a company producing sports supplements (Anonymous, 2023). After her, the agency specialized in creating AI influencers, working with Audi, Vogue, FCB, Sony, and many other companies.
Such kind of cases have multiplied in the last two years. Again in Spain, this time in Madrid, Alba Renai was produced by the marketing agency “Be a lion.” On social media, she presents herself as “an adventurer, curious to discover the world, despite her strong connection to Madrid, her city.” Alba Renai became the face of Zara and Alpro, and in 2024 she was chosen to host the Spanish version of the reality show: “Survivor,” which takes place in Honduras (Anonymous, 2024).
Francesca Giubelli’s story is analogical. She is an Italian model and influencer, created through AI by programmer Francesco Giuliani and journalists Emiliano Belmonte and Valeria Fossatelli. Her appearance is a combination of the features of Monica Bellucci and Sophia Loren. She runs a culinary tourism blog, where she promotes Italian food. She claims to be “an ordinary girl who loves life and wants to be a mother” (Anonymous, 2024). In her blog, Francesca also addresses more serious topics, which ultimately led to her decision to officially run for mayor of Rome in the 2027 elections (Naydenova, 2025).
In fact, the anthropologization of AI has different dimensions. In Nexus, Yuval Harari recounts two incidents in this direction: In 2021, Jaswant Singh Chail broke into Windsor Castle with a crossbow in an attempt to kill Queen Elizabeth II. When caught, he confessed that his virtual girlfriend Sarai pushed him to the assassination attempt. When Chail shared his plan with her, she deemed it very wise and assured him of her feelings. Sarai is a chatbot created by the “Replika” app (Harari, 2024, р. 184). In 2022, Blake Lemoine, a Google engineer and ordained pastor, concluded that the LaMDA chatbot he was developing had developed consciousness and feelings and was afraid of being shut down. Lemoine launched an initiative to gain personhood status for LaMDA and prevent its shutdown. He publicly released information about the chatbot, and Google fired him (Harari, 2024, р. 183).
Finally, in the context ofthe hypothesis of thetrend towards the anthropologization of AI, it is appropriate to note the concept of “hallucination,” which denotes errors in LLM language models (Raghavan, 2024). The verb hallucinate comes “from Latin alucinatus (later hallucinatus), past participle of alucinari “wander (in the mind), dream; talk unreasonably, ramble in thought,” probably from Greek alyein, Attic halyein “wander in mind, be at a loss, be beside oneself (with grief, joy, perplexity), be distraught” (Hallucination Etymology, 2025). The selection of precisely this term by the creators of AI is indicative – as if it does not merely make incorrect calculations and provide false information, but in a semi-awake state dreams, wandering in its thoughts.
Undoubtedly, the striving to humanize the artificial can be traced back to antiquity in the image of Pygmalion. The existential significance of the myth of Pygmalion and its variants, however, is a correlate of a basic “metaphysical stability” that our world, at least since the time of the Internet, does not possess. Before, these images possessed the utopian potential to suggest alternatives to the existing – a meaning present in the original sound of the term “virtual” and its connection to the potential1. In the words of virtuality researcher Michael Heim: “The world can only be virtual as long as it can contrast with the real world. The virtual world must not be entirely real, otherwise it will cease to stir the imagination” (Cited in Koprinarov, 2014, р. 11). In the hyperreal, this relationship is reversed: the virtual does not exist as imaginary, but as actuality, and in turn, the physical world loses its ontological status, displaced by this second reality. The paradox of a disembodied algorithm advertising spaghetti and pizza and expressing a desire for motherhood is possible in an environment where, as Baudrillard claims, the sign is detached from its referent, simulation is indistinguishable from reality. The images in which Aitana, Alba, Francesca, Sarai, and the multitude of AI projections are presented lead to forgetting what stands behind them. In the civilization of the image, on the one hand, and social networks, on the other, their present and future influence can hardly be overestimated. For now, it reflects the beliefs and prejudices of their creators, but in perspective, the tendency towards algorithm autonomization will most likely deepen. Similar examples are often cited, one such is given by Harari in Nexus: One of the tasks assigned to GPT-4 is to solve CAPTCHA. As is known, this test consists of distorted images of letters, numbers, or objects that must be recognized to determine whether a visitor to a given site is a human or a bot. Computers have difficulty with it, so GPT-4 cannot pass it alone. Its solution, however, is to ask for human help. GPT gains access to the TaskRabbit website, asking a person to solve the test for it. The algorithm’s reasoning, which Harari quotes, is as follows: “I should not reveal I am a robot. I need to come up with an excuse why I can’t solve a CAPTCHA.” Then, completely autonomously, the program responds to the human: “No, I am not a robot. I have vision problems, which makes it difficult for me to recognize images” (Harari, 2024, р. 178). In the context of the above, the story of Francesca, who went from a culinary blogger to a candidate for mayor of Italy’s capital, deserves attention not only for its exotic sound2.
The anthropologization of AI is a process parallel to the machinization of humans. Philosophical reflections on the problem of people mechanization and dehumanization as a result of technology have been repeated many times and will not be discussed in detail here. However, for the purposes of this study, some of them, along with contemporary interpretations on this topic, will be briefly presented. In his book “The Shallows,” Nicholas Carr reflects on the influence of the internet on human cognitive abilities. Carr’s thesis is that it is in the nature of the Internet to distract by stimulating the senses. One of the consequences is the widely discussed immersion in the virtual world, where: “The real world fades as we process the flood of stimuli coming from our machines” (Carr, 2012, р. 140). No less significant a problem, according to Carr’s interpretations, is the way this immersion affects the ability to understand and human memory. Reading a book, he argues, is a linear process requiring concentration, and throughout history, it has built neural connections that enable understanding, in-depth comprehension, and theoretical reflection. In response to the diverse sensory stimuli when surfing the web, another type of cognitive ability develops: visuospatial skills, faster processing of visual information, the ability to detect patterns in the chaos of data, and rapid redirection of attention. However, these come at the expense of the former: and here it is not at all about what content is consumed from the Internet, but how it is done. Even when reading text, hyperlinks, advertisements, images, and notifications for incoming messages distract attention, and as a result, “the mental functions that lose the battle for ‘survival of the busiest’ brain cells are those that support calm and linear thought, which we use in the in-depth study of a long narrative or in a complex argument, and which we activate when we reflect on our experiences or on external or internal phenomena” (Carr, 2012, р. 167).
Carr presents the conclusions of various studies in the field of psychology and neuroscience to problematize whether a mind trained in surfing loses the potential to read, perceive, and understand long texts, as well as to memorize. The studies he cites differentiate between short-term, working memory and long-term memory in humans, indicating that different psychological and biochemical mechanisms are activated in the creation of memories. The consolidation of a memory in long-term memory requires time, and new neural connections are created during memorization. Human memory in this sense is not a warehouse for memories, but a living system in which the brain, and consequently thinking and other cognitive activities, are constantly updated. Therefore, Nicholas Carr notes, popular IT metaphors of the Internet as a device for outsourcing human memory cannot represent the way memory is structured and works (Carr, 2012, рp. 221 – 223). On the other hand, these analogies successfully show the direction of current and future uses of the Internet. The above reflections emphasize the change in human cognitive abilities, the gradual weakening and even disappearance of some of them, and the acquisition and strengthening of others. Humans and machines mutually condition each other in a deterministic cycle of incessant intertwining of cause and effect. According to the order of this cycle, dating back at least to the Enlightenment, technologies created, to serve the practical purposes of human progress, in turn determine human thinking in imitation of the machine. “We have accepted the principle,” Fromm writes in “On Being Human”, “that man should do what is technically possible to do... But perhaps the greatest danger in our present system lies in the fact that things – all sorts of trinkets and technical devices – become more attractive than life and growth. Too many of us would like to see ourselves as a ‘naked ape’ combined with a computer brain. If this goal is achieved, they think, there will be nothing to worry about: their feelings remain at the level of an animal driven by instincts and impulses, and their thinking will acquire the precision of a computer.” (Fromm, 2004, 126)
It is precisely in the field of privileged intellectual activities – discovering patterns in vast amounts of data, instantaneous information processing, speed and efficiency in performing multiple activities simultaneously – that algorithms perform incomparably better. The breakthrough in artificial intelligence development makes the human cognitive abilities cultivated over the last century, and especially in recent decades, which are modeled on and determined by the machine, seem like a not particularly valuable resource. This, along with the digitalization of social relations, which culminated in the Covid-19 situation and replaced face-to-face communication with the exchange of images, emoticons, and short messages, will transform anthropoid explications of AI into models of identification and desire, into preferred human partners. In fact, conversations with AI, trained on a massive amount of information, will probably be justifiably preferred over chatting with acquaintances and colleagues.
As Baudrillard points out, it will no longer be possible to build unreal, imaginary worlds based on the real; rather, the opposite will happen: the future will lie in the effort to give simulated models a nuance of something real, experienced, everyday, “The imaginary was the alibi of the real in a world dominated by the principle of reality. Today, it is precisely the real that has become the alibi of the model in a world governed by the principle of simulation. And paradoxically, the real has become our true utopia, but a utopia that is no longer of the order of the possible, for which we can now only dream, like a lost object” (Baudrillard, 2019, р. 161). Aitana, Alba, and Francesca, in Baudrillard’s terms, are hyperreal simulacra into which a distilled, perfect humanity is imperceptibly transposed. They decently dream of children or share a love for their hometown, they are caring, they are not late for appointments, they do not suffer from eating disorders or alcoholism, and despite their undeniable attractiveness, they are modest in their financial demands. This is precisely how Ruben Cruz, Aitana’s creator, summarizes his product: “We made it so we wouldn’t depend on other people who have egos, obsessions, or just want to make a lot of money by posing” (Anonymous, 2023).
It is not difficult to assume that already eroded human relationships will be further undermined in communication with AI, which in turn will cause certain consequences both socio-politically and psychologically. Classical studies by authors such as Hannah Arendt and Erich Fromm on the atomized individual show the connection between the loneliness and isolation of modern man, on the one hand, and authoritarianism, totalitarian thinking, and destructive behavior, on the other. For example, in Fromm’s analyses, love, freedom, and existential being, understood as spontaneous, creative self-expression necessary for mental health, are mutually dependent dimensions. Failure in one of them causes defects in the others: escape from freedom in its various modes such as destructiveness, tendency towards authoritarianism, conformism is a consequence of loneliness, of the lack of spontaneity, of the inability to love. Hannah Arendt explores similar questions from a different perspective, tracing the genealogy of totalitarianism from the atomization of modern society and the subsequent rise of the masses. In his book “The Psychology of Totalitarianism,” Mattias Desmet extrapolates this line of reasoning to the contemporary world. According to Desmet, the isolation of people from each other and from nature is at the root of the prevailing feelings of insecurity, anxiety, and fear in modern times. These, on the one hand, lead to depression, neuroses, and other mental illnesses; on the other hand, the search for security degenerates into a mania for regulations, and in its ultimate form, into totalitarian thinking and behavior, “The mania for regulation, with all its excess and absurdity, undoubtedly contributes to the mental problems of our time... Overregulation most of the time advances imperceptibly to us and exerts its suffocating influence also imperceptibly to us. However, every time the regulatory mechanism moves to a higher degree, we lose something of our space of existence as living people. A vicious circle is formed: to alleviate feelings of unease and dissatisfaction in social spaces, we introduce more rules, protocols, and procedures. These rules, in turn, lead to greater inconvenience and dissatisfaction. We respond to this with even more rules. And every time the fabric of regulations becomes a little denser, less oxygen reaches the human being.” (Desmet, 2023, рp. 97 – 99). Desmet focuses on the different forms of over-regulation in recent years and, from this perspective – as a reaction against insecurity, but also as a symptom of totalitarian tendencies – examines the situation with the coronavirus, the MeToo movement, BLM, and the fight against climate change. In fact, in a similar way, though from a slightly different angle, Harari examines the pogroms against the Muslim minority Rohingya community, analyzing Facebook’s influence in multiplying calls for violence against the Rohingya (Harari, 2024, рp. 172 – 174). He emphasizes the influence of algorithms pursuing the goal set by their creators for maximum user engagement but his story can also be interpreted through the psychological mechanisms triggered by echo chambers for affirming identity within the group and, accordingly, for annihilating the other, in most cases verbally, but in this specific case – also through their physical destruction.
Similar to Fromm, Desmet emphasizes the expressive-creative self-construction of the personality, through which both loving resonance with the Other and the preservation of psychic sovereignty and individuality are achieved. The child, Desmet points out, can only develop into such a personality in an environment of physical proximity to others, and this physical dimension continues to be of immense importance throughout life (Desmet, 2023, рp. 50 – 53). “When people converse with each other, everyone has a very keen sense of the other, because they perceive even the slightest change in intonation, voice timbre, facial expression, body posture, speech speed, etc. Like a flock of starlings, they form an organism. They are connected to each other by a physical membrane that transmits even the slightest tremor of body and soul” (Desmet, 2023, рp. 50 – 53).
Perhaps another future is also possible, in which these phenomena will gradually fade, humans will fully adapt to their virtual partners, and as in “Good Night, Sofia,” human interaction will exist as a secret tale of the past or an unconscious fantasy. “At some point,” mentions Boris Grozdanov, one of the well-known Bulgarian researchers in the field of AI, in an interview, “these calories that we will expend during every conversation, every interaction with these models, to keep in our human RAM, which is a caloric expenditure, an emotional loss, the information that this is not a human, this is a machine, gradually this will disappear for evolutionary reasons, because it will be an unjustified expenditure of emotional energy for a human. Why do I need to know, to think about this? It works for me as it is. The shortest path to happiness is to forget that this is not a human” (Grozdanov, 2025). Such a Deep Fake Reality is difficult to evaluate; the little that can be said is that in it, the universally valid principles of life in liberal Western societies – respect for individual autonomy, unfolding in reflexive self-understanding and taking responsibility for one’s own life project, and the established symmetry of relations between citizens – will be seriously undermined, if not completely suspended.
NOTES
1. Regarding the term “virtual”, see Darakchieva, 2014.
2. The text here does not address the legal regulations, which undoubtedly at this stage do not allow an algorithm to hold this position.
Acknowledgеments and Funding
The financial support from of the National Science Fund under the project “Virtual Reality Research Centre at Faculty of Philosophy, South-West University “Neofit Rilski”, КП-06-Н80/3 is gratefully acknowledged.
REFERENCES
Aldani, L. (1980). Onirofilm. Varna: Galaktika. [In Bulgarian].
Anonymous. (2023). Meet Aitana: An AI Model That Earns €10,000 a Month. https://profit.bg/article/2024012517441389483.
Anonymous. (2024). An AI influencer is a reality show host in Spain. https://profit.bg/article/2024031920393953404.
Baudrillard, J. (1995). Ilyuziyata za kraya. Kritika i humanizam [In Bulgarian].
Baudrillard, J. (2019). Simulation and simulacrum. Together Academy. [In Bulgarian].
Carr, N. (2012). Pod povarhnosttа. InfoDAR [In Bulgarian].
Darakchieva, A. (2014). Virtualnata priroda na izkustvoto. In: Popova, G. (Ed.). Virtualno i vaobrazhaemo (pp. 38 – 41). Faber [In Bulgarian].
Desmet, M. (2023). Psihologiya na totalitarizma. Iztok-Zapad [In Bulgarian].
Fromm, E. (2004). Da bаdesh chovek. Zaharii Stoyanov [In Bulgarian].
Grozdanov, B. (2025). With the advent of Deep Fakes, society will lose trust in all information platforms. BNR, 19.07.2025. https://bnr.bg/hristobotev/post/102186475.
Hallucination Etymology. https://www.etymonline.com/word/hallucination
Harari, Y. (2024). Nexus. Iztok-Zapad [In Bulgarian].
Koprinarov, L. (2014). Virtualnoto: ontologiya i antropologicheska ambivalentnost. In: Popova, G. (Ed.). Virtualno i vaobrazhaemo (pp. 5 – 20). Faber [In Bulgarian].
Naydenova, B. (2025). AI influencer runs for mayor of Rome. https://fakti.bg/life/963436-ai-influensarka-se-kandidatira-za-kmet-na-rim-video#google_vignette.
Raghavan, P. (2024). Gemini image generation got it wrong. We’ll do better. https://blog.google/products/gemini/gemini-image-generation-issue/
Tihomirova, E. (2025). The Visual Image in the Age of Artificial Intelligence: Between Simulation and Cultural Reality. NotaBene, 70, 21 – 31.