Political deepfakes have prompted outcry about the diminishing trustworthiness of visual depictions, and the epistemic and political threat this poses. Yet this new technique is being used overwhelmingly to create pornography, raising the question of what, if anything, is wrong with the creation of deepfake pornography. Traditional objections focusing on the sexual abuse of those depicted fail to apply to deepfakes. Other objections—that the use and consumption of pornography harms the viewer or other (non-depicted) individuals—fail to explain the objection (...) that a depicted person might have to the creation of deepfake pornography that utilises images of them. My argument offers such an explanation. It begins by noting that the creation of sexual images requires an act of consent, separate from any consent needed for the acts depicted. Once we have separated these out, we can see that a demand for consent can arise when a sexual image is of us, even when no sexual activity was actually engaged in, as in the case of deepfake pornography. I then demonstrate that there are two ways in which an image can be ‘of us’, both of which can exist in the case of deepfakes. Thus, I argue: if a person, their likeness, or their photograph is used to create pornography, their consent is required. Whenever the person depicted does not consent (or in the case of children, can’t consent), that person is wronged by the creation of deepfake pornography and has a claim against its production. (shrink)

Deepfakes are extremely realistic audio/video media. They are produced via a complex machine-learning process, one that centrally involves training an algorithm on hundreds or thousands of audio/video recordings of an object or person, S, with the aim of either creating entirely new audio/video media of S or else altering existing audio/video media of S. Deepfakes are widely predicted to have deleterious consequences (principally, moral and epistemic ones) for both individuals and various of our social practices and institutions. In this introduction (...) to the Topical Collection, I first survey existing philosophical research on deepfakes (Sects. 2 and 3). I then give an overview of the papers that comprise the Collection (Sect. 4). Finally, I conclude with remarks on a line of argument made in a number of papers in the Topical Collection: that deepfakes may cause their own demise (Sect. 5). (shrink)

Deepfakes are AI-generated media. When produced competently, they are near-indistinguishable from genuine recordings and so may mislead viewers about the actions of the individuals they depict. For this reason, it is thought to be only a matter of time before deepfakes have deleterious consequences for democratic procedures, elections in particular. But this pessimistic view about deepfakes and their relation to democracy is flawed, whether it means to pick out current deepfakes or future ones. Rather than advocating for an optimistic view (...) in its place, I outline the opposite: a nihilistic account of deepfakes and their relation to democracy. On the nihilistic view, the harms that deepfakes pose for democracy are significantly more serious than those implied by the pessimistic view. Nihilism says that the real threat that deepfakes pose for democracy is that their existence counts against reforming current politics to be more truth-oriented. (shrink)

Deepfake technology uses machine learning to fabricate video and audio recordings that represent people doing and saying things they've never done. In coming years, malicious actors will likely use this technology in attempts to manipulate public discourse. This paper prepares for that danger by explicating the unappreciated way in which recordings have so far provided an epistemic backstop to our testimonial practices. Our reasonable trust in the testimony of others depends, to a surprising extent, on the regulative effects of (...) the ever-present possibility of recordings of the events they testify about. As deepfakes erode the epistemic value of recordings, we may then face an even more consequential challenge to the reliability of our testimonial practices themselves. (shrink)

The recent emergence of machine-manipulated media raises an important societal question: How can we know whether a video that we watch is real or fake? In two online studies with 15,016 participants, we present authentic videos and deepfakes and ask participants to identify which is which. We compare the performance of ordinary human observers with the leading computer vision deepfake detection model and find them similarly accurate, while making different kinds of mistakes. Together, participants with access to the model’s (...) prediction are more accurate than either alone, but inaccurate model predictions often decrease participants’ accuracy. To probe the relative strengths and weaknesses of humans and machines as detectors of deepfakes, we examine human and machine performance across video-level features, and we evaluate the impact of preregistered randomized interventions on deepfake detection. We find that manipulations designed to disrupt visual processing of faces hinder human participants’ performance while mostly not affecting the model’s performance, suggesting a role for specialized cognitive capacities in explaining human deepfake detection performance. -/- . (shrink)

The recent proliferation of deepfakes and other digitally produced deceptive representations has revived the debate on the epistemic robustness of photography and other mechanically produced images. Authors such as Rini (2020) and Fallis (2021) claim that the proliferation of deepfakes pose a serious threat to the reliability and the epistemic value of photographs and videos. In particular, Fallis adopts a Skyrmsian account of how signals carry information (Skyrms, 2010) to argue that the existence of deepfakes significantly reduces the information that (...) images carry about the world, which undermines their reliability as a source of evidence. In this paper, we focus on Fallis’ version of the challenge, but our results can be generalized to address similar pessimistic views such as Rini’s. More generally, we offer an account of the epistemic robustness of photography and videos that allows us to understand these systems of representation as continuous with other means of information transmission we find in nature. This account will then give us the necessary tools to put Fallis’ claims into perspective: using a richer approach to animal signaling based on the signaling model of communication (Maynard-Smith and Harper, 2003), we will claim that, while it might be true that deepfake technology increases the probability of obtaining false positives, the dimension of the epistemic threat involved might still be negligible. (shrink)

Deepfakes are algorithmically modified video and audio recordings that project one person’s appearance on to that of another, creating an apparent recording of an event that never took place. Many scholars and journalists have begun attending to the political risks of deepfake deception. Here we investigate other ways in which deepfakes have the potential to cause deeper harms than have been appreciated. First, we consider a form of objectification that occurs in deepfaked ‘frankenporn’ that digitally fuses the parts of (...) different women to create pliable characters incapable of giving consent to their depiction. Next, we develop the idea of ‘illocutionary wronging’, in which an individual is forced to engage in speech acts they would prefer to avoid in order to deny or correct the misleading evidence of a publicized deepfake. Finally, we consider the risk that deepfakes may facilitate campaigns of ‘panoptic gaslighting’, where many systematically altered recordings of a single person's life undermine their memory, eroding their sense of self and ability to engage with others. Taken together, these harms illustrate the roles that social epistemology and technological vulnerabilities play in human ethical life. (shrink)

Deepfake technology, powered by artificial intelligence and deep learning, has rapidly advanced, enabling the creation of highly realistic synthetic media. While it presents opportunities in entertainment and creative applications, deepfakes pose significant risks, including misinformation, identity fraud, and threats to privacy and national security. This study explores the evolution of deepfake technology, its implications, and current detection techniques. Existing methods for deepfake detection, such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), and generative adversarial networks (GANs), (...) are examined, highlighting their effectiveness and limitations. The study also reviews state-of-the-art approaches in image forensics, phoneme-viseme mismatch detection, and adversarial training to counter deepfake threats. Moreover, the ethical and legal challenges surrounding deepfakes are discussed, emphasizing the need for policy regulations and collaborative efforts between governments, tech companies, and researchers. As deepfake technology continues to evolve, so must detection strategies, integrating multimodal analysis and real-time verification systems. This research underscores the importance of developing robust detection frameworks and public awareness initiatives to mitigate the risks associated with deepfakes. Future directions include enhancing detection algorithms through explainable AI, improving dataset quality, and integrating blockchain for digital content authentication. By providing a comprehensive analysis of deepfake creation, detection, and countermeasures, this study contributes to the ongoing discourse on synthetic media and its societal impact. Addressing these challenges requires interdisciplinary collaboration and continuous innovation to safeguard digital integrity and trust in the information ecosystem. (shrink)

[Author note: There is a video explainer of this paper on youtube at the new work in philosophy channel (search for surname+deepfakes).] -/- It is widely thought that deepfake videos are a significant and unprecedented threat to our epistemic practices. In some writing about deepfakes, manipulated videos appear as the harbingers of an unprecedented _epistemic apocalypse_. In this paper I want to take a critical look at some of the more catastrophic predictions about deepfake videos. I will argue (...) for three claims: (1) that once we recognise the role of social norms in the epistemology of recordings, deepfakes are much less concerning, (2) that the history of photographic manipulation reveals some important precedents, correcting claims about the novelty of deepfakes, and (3) that proposed solutions to deepfakes have been overly focused on technological interventions. My overall goal is not so much to argue that deepfakes are not a problem, but to argue that behind concerns around deepfakes lie a more general class of social problems about the organisation of our epistemic practices. (shrink)

Deepfake technology can be used to produce videos of real individuals, saying and doing things that they never in fact said or did, that appear highly authentic. Having accepted the premise that Deepfake content can constitute a legitimate form of expression, it is not immediately clear where the rights of content producers and distributors end, and where the rights of individuals whose likenesses are used in this content begin. This paper explores the question of whether it can be (...) plausibly argued that Deepfake content involving the likenesses of real individuals violates the rights of these individuals. (shrink)

This paper takes up the question of how videographic public announcements (VPAs)---i.e. videos that a wide swath of the public sees and knows that everyone else can see too--- have functioned to mobilize people politically, and how the presence of deepfakes in our information environment stands to change the dynamics of this mobilization. Existing work by Regina Rini, Don Fallis and others has focused on the ways that deepfakes might interrupt our acquisition of first-order knowledge through videos. But I point (...) out that even where every audience member takes a video to be veridical, where first-order knowledge acquisition is secure, an audience aware of deepfakes in their environment will not acquire higher order knowledge in the way that has erstwhile been characteristic for audiences of VPAs. Engaging with ideas from the literatures on public announcement logic, common knowledge, and convention, I enumerate a variety of ways in which we should expect this absence of higher order knowledge to throw up barriers to political mobilization. I go on to apply my analysis of VPAs to the mechanisms by which an uptick in publicly available videos of police brutality over the last decade, mediated by camera phones and social media, was responsible for the largest mass protests in US history in summer 2020. This makes vivid the stakes of the transformation in our mobilizing environment that I've claimed deepfakes effect: where we lose the common-knowledge-generating effects of VPAs, there are fresh obstacles to this sort of mass mobilization. (shrink)

Recent develops in AI technology have led to increasingly sophisticated forms of video manipulation. One such form has been the advent of deepfakes. Deepfakes are AI-generated videos that typically depict people doing and saying things they never did. In this paper, I demonstrate that there is a close structural relationship between deepfakes and more traditional fake barn cases in epistemology. Specifically, I argue that deepfakes generate an analogous degree of epistemic risk to that which is found in traditional cases. Given (...) that barn cases have posed a long-standing challenge for virtue-theoretic accounts of knowledge, I consider whether a similar challenge extends to deepfakes. In doing so, I consider how Duncan Pritchard’s recent anti-risk virtue epistemology meets the challenge. While Pritchard’s account avoids problems in traditional barn cases, I claim that it leads to local scepticism about knowledge from online videos in the case of deepfakes. I end by considering how two alternative virtue-theoretic approaches might vindicate our epistemic dependence on videos in an increasingly digital world. (shrink)

In recent years, a number of philosophers have turned their attention to developments in Artificial Intelligence, and in particular to deepfakes. A deepfake is a portmanteau of ‘deep learning' and ‘fake', and for the most part they are videos which depict people doing and saying things they never did. As a result, much of the emerging literature on deepfakes has turned on questions of trust, harms, and information-sharing. In this paper, I add to the emerging concerns around deepfakes by (...) drawing on resources from vice epistemology. As deepfakes become more sophisticated, I claim, they will develop to be a source of online epistemic corruption. More specifically, they will encourage consumers of digital online media to cultivate and manifest various epistemic vices. My immediate focus in this paper is on their propensity to encourage the development of what I call ‘intellectual cynicism'. After sketching a rough account of this epistemic vice, I go on to suggest that we can partially offset such cynicism – and fears around deceptive online media more generally – by encouraging the development what I term a trained ‘digital sensibility'. This, I contend, involves a calibrated sensitivity to the epistemic merits of online content. (shrink)

Could 2024 be the year heralding what one could term the deepfake universe apocalypse scenario or could it be the year that a future history of science may e.g. interpret as the year of the first literally universe-sized algorithmic hype bubble? This commentary introduces the metaphor of "GPT-Universe" and the assumptions hidden beneath it.

Would suppressing deepfakes violate freedom of expression norms? The question is pressing because the deepfake phenomenon in its more poisonous manifestations appears to call for a response, and automated targeting of some kind looks to be the most practically viable. Two simple answers are rejected: that deepfakes do not deserve protection under freedom of expression legislation because they are fake by definition; and that deepfakes can be targeted if but only if they are misleadingly presented as authentic. To make (...) progress, following a discussion of why freedom of expression deserves protection in a democracy, the question is reframed. At issue is not whether the arrival of deepfakes brings new and potentially serious dangers (it does), nor whether these dangers call for measures that potentially limit freedom of expression (they do), but whether the need for such measures raises any new and unfamiliar freedom-of-expression challenges. The answer to that question, surprisingly, is no. The balancing act needed to cope with the arrival of deepfakes brings plenty of difficulties, certainly, but none of the measures likely to be effective in tackling deepfake harms raises freedom-of-expression concerns that aren’t familiar from consideration of non-deepfake harms. In that respect, at least, the arrival of deepfakes makes no difference. (shrink)

Deepfake technology has quickly become one of the most disruptive new technologies in the digital information ecosystem. It can make convincing audio-visual content that looks just like real media. Although much of the current research has looked at deepfakes from legal, technological, or ethical points of view, this article uses a psychological point of view to explain their particular potency. The main point is that deepfakes don't work mostly as rational lies; instead, they work as cognitive exploits, types of (...) "perceptual malware" that attack the way people trust each other. The article illustrates how deepfakes exploit entrenched biases, such as the truth bias, familiarity effects, and the fluency of perceptual processing, by referencing cognitive psychology and neuroscience. In doing so, they avoid higher-order logical inspection and use autonomic processes of recognition and belief construction. The paper also puts deepfakes in the context of other research on false information and how memories can change, suggesting that fake media can create false memories or strengthen existing biases through repeated exposure. Deepfakes undermine epistemic trust by eroding the dependability of visual and aural evidence, which have historically been regarded as the "gold standard" of credibility, in contrast to conventional kinds of propaganda. This situation not only puts democratic discourse and trust in the media at risk, but it also poses a serious psychological challenge: fighting deepfakes requires using cognitive and perceptual mechanisms that are mostly outside of conscious control. This article provides an interdisciplinary framework for comprehending deepfakes as psychological weapons by integrating insights from cognitive science, information warfare studies, and emerging media research. It suggests that protecting against this phenomenon necessitates treatments that transcend detection technologies, highlighting cognitive inoculation, media literacy, and the development of "epistemic resilience" in an age where visual evidence is increasingly distrusted. (shrink)

While the concerns about media trust have historical roots, the dynamics have evolved with technological advancements, the rise of digital media, and the challenges posed by disinformation in the 21st century. Deepfake – the media content (text, images, videos) created by AI technology – is just another concern for journalists, politicians and society in the current media trust crisis. It is a tool that can be successfully applied in an information warfare. To be honest, face and voice spoofing has (...) always existed, but deepfake technology may now make fake videos faster, cheaper, with high quality and is accessible to anybody. Although deepfake videos have the potential to undermine trust in journalism, their actual impact on audiences remains limited due to information overload, increased suspicion and skeptical attitude, decreased trust in media and advanced audience adaptability for orientation in the information flow. (shrink)

Der Begriff Deepfake ist zum Symbol digitaler Täuschung geworden. Doch dieselbe Technologie, die Fiktion erzeugt, eröffnet zugleich neue Räume der Wahrhaftigkeit. Dieser Essay schlägt einen Perspektivwechsel vor: weg von der Angst vor Simulation – hin zu einer Ethik der Absicht. Aufbauend auf dem Reiter-Framework for Ethical Synthesis wird das Konzept der DeepCreation vorgestellt: eine Praxis, in der KI nicht lügt, sondern ausdrückt. Ziel ist eine neue Form digitaler Glaubwürdigkeit, in der Technik und Mensch eine gemeinsame ästhetische Verantwortung tragen.

Non-Consensual Sexual Deepfakes (NCSD) are a growing issue of our onlife experience. We argue that, far from harming individuals merely indirectly, NCSD constitute a _Direct Personal Harm_ to the persons they depict, because they can be understood as parts of their personal identity. We defend this claim by developing a narrative theory of identity according to which a person is constituted by their life story—a story that is not solely self-authored but socially shaped and often beyond one’s control. On this (...) basis, we argue that digital artifacts such as NCSD can become elements of a person’s narrative identity. We ground our account in a four-dimensionalist metaphysics that treats persons as aggregates of both biological and digital temporal parts. Finally, by examining the verbal behavior of imaginary and actual victims of NCSD, we show that this account is less counterintuitive than it may initially appear and helps make sense of why many experience NCSD as violations of their person, rather than merely of their representation. (shrink)

: Deepfake technology enables the creation of hyper-realistic fake videos, posing significant threats in domains like politics, law enforcement, and cybersecurity. This project proposes a deepfake detection framework leveraging facial feature embeddings using FaceNet512, followed by classification through transfer learning models. Unlike traditional CNN-based detectors that analyze full video frames, this method isolates and processes facial data, ensuring higher efficiency and accuracy. Additionally, a real-time email alert system notifies users upon deepfake detection. Evaluation across benchmark datasets like (...) DFDC and FaceForensics++ demonstrates high accuracy, robustness, and suitability for real-time deployment. The proposed system is designed to detect subtle facial manipulations using efficient feature extraction and classification techniques. By focusing on key facial landmarks and minimizing the amount of data analyzed, it ensures quick and effective identification of forged content. The use of pre-trained models via transfer learning significantly reduces the training time while maintaining high generalization across various datasets. Integration with email alerts also provides timely responses for practical security applications. The solution is versatile, scalable, and offers real - world utility in environments like social media platforms, digital forensics, and content authentication. It offers a balanced approach to performance and usability, setting a foundation for future improvements like audio detection, multimodal input support, and deployment on mobile or edge device. (shrink)

Deepfakes pose a multi-faceted threat to the acquisition of knowledge. It is widely hoped that technological solutions—in the form of artificially intelligent systems for detecting deepfakes—will help to address this threat. I argue that the prospects for purely technological solutions to the problem of deepfakes are dim. Especially given the evolving nature of the threat, technological solutions cannot be expected to prevent deception at the hands of deepfakes, or to preserve the authority of video footage. Moreover, the success of such (...) technologies depends on institutional trust that is in short supply. Finally, outsourcing the discrimination between the real and the fake to automated, largely opaque systems runs the risk of undermining epistemic autonomy. (shrink)

AI, especially in the case of Deepfakes, has the capacity to undermine our confidence in the original, genuine, authentic nature of what we see and hear. And yet digital technologies, in the form of databases and other detection tools also make it easier to spot forgeries and to establish the authenticity of a work. Using the notion of ectypes, this paper discusses current conceptions of authenticity and reproduction and examines how, in the future, these might be adapted for use in (...) the digital sphere. (shrink)

This epistemic project examines an understudied existential risk emerging in the deepfake era: the fortunately up to this time (but not indefinitely so) reversible peril of humanity’s epistemic self-sabotage through an overestimation of algorithms linked to quantitative aspects and a paired underestimation of the own epistemic potential whose manifestations are in principle expressible via scientifically analyzable but currently often neglected qualitative facets. This scenario is metaphorically referred to as "π-Doom scenario". Instead of carefully crafting opaque hypotheses and formulating probabilistic (...) predictions of timelines that could easily evade scrutiny and/or could even themselves be initially generated by algorithms, scientifically and responsibly counteracting π-Doom requires new better riskier explanatory theories of intelligence, creativity and consciousness. A brief non-comprehensive literature review suggests that quite a few scientific instantiations (all of which have not yet been provisionally refuted) of that transformative requirement are already available at present. In short, π-Doom – the voluntarily or unintentionally effectuated cognitive resignation engendering an irrational algorithmically-motored process of running around in (epistemic) circles that could endanger the immediate future of a vulnerable civilization like present-day humanity – is wholly contingent on a choice of focus. (shrink)

In this paper, I draw a distinction between two types of deepfake, and unpack the deceptive strategies that are made possible by the second. The first category, which has been the focus of existing literature on the topic, consists of those deepfakes that act as a fabricated record of events, talk, and action, where any utterances included in the footage are not addressed to the audience of the deepfake. For instance, a fake video of two politicians conversing with (...) one another. The second category consists of those deepfakes that direct an illocutionary speech act - such as a request, injunction, invitation, or promise - to an addressee who is located outside of the recording. For instance, fake footage of a company director instructing their employee to make a payment, or of a military official urging the populace to flee for safety. Whereas the former category may deceive an audience by giving rise to false beliefs, the latter can more directly manipulate an agent's actions: the speech act's addressee may be moved to accept an invitation or a summons, follow a command, or heed a warning, and in doing so further a deceiver's unethical ends. (shrink)

The rapid development of generative artificial intelligence has introduced deepfakes as a new and severe threat to personality rights. Synthetic manipulation of voice, image and identity enables the creation of highly realistic fabricated content capable of inflicting immediate and substantial harm on human dignity. Due to their inherently offensive nature, this study argues that moral damages resulting from deepfakes must be presumed — constituting a clear case of “in re ipsa” damage — without requiring proof of psychological suffering or concrete (...) injury. The spread of such material is worsened by the algorithmic dynamics of digital platforms, which privilege sensational content and often profit from fraudulent advertisements, as recent investigations have shown. Moreover, deepfakes are used in contexts ranging from non-consensual synthetic pornography to financial scams and political misinformation, revealing their multifaceted and severe harmful potential. Given this scenario, civil liability — of both creators and digital platforms — must be reinforced and may even assume strict (objective) liability due to the inherent risks of the digital environment. Therefore, a renewed legal and jurisprudential approach is necessary to ensure the effective protection of personality rights. (shrink)

In January 2024, non-consensual deepfakes came to public attention with the spread of AI generated sexually abusive images of Taylor Swift. Although this brought new found energy to the debate on what some call non-consensual synthetic intimate images (i.e. images that use technology such as AI or photoshop to make sexual images of a person without their consent), female celebrities like Swift have had deepfakes like these made of them for years. In 2017, a Reddit user named “deepfakes” posted several (...) videos in which he had used opensource machine learning tools to swap the faces of female celebrities on to the faces of female porn actors, displaying what appeared to be live video footage of the celebrity engaging in sex acts she never engaged in. Since that time, deepfake technology has advanced astronomically. What once were choppy sexualized videos are now nearly flawless videos that can be difficult to distinguish from a real video. According to recent research on deepfakes by Sensity AI, this technology has been used primarily on women to create sexual videos. These women’s sexual autonomy has been co-opted for the purposed of gratifying men’s sexual pleasure, but have also been used in campaigns to delegitimize and humiliate female journalists and politicians. In Canada, civil and criminal legislation has addressed the non-consensual distribution of intimate images, but in only a few provinces – British Columbia, New Brunswick, Prince Edward Island and Saskatchewan – does this offence include altered images that could include deepfakes. This paper explores the evolution of synthetic technology such as AI image generators and Canada’s legal responses to the non-consensual sharing of intimate images. (shrink)

Everyday (mis)uses of deepfakes define prevailing conceptualizations of what they are and the moral stakes in their deployment. But one complication in understanding deepfakes is that they are not photographic yet nonetheless manipulate lens-based recordings with the intent of mimicking photographs. The harmfulness of deepfakes, moreover, significantly depends on their potential to be mistaken for photographs and on the belief that photographs capture actual events, a tenet known as the transparency thesis, which scholars have somewhat ironically attacked by citing digital (...) imaging techniques as counterexamples. Combining these positions, this paper sets out two core points: (1) that conceptions about the nature of photography introduce imperatives about its uses; and (2) that popular cultural understandings of photography imply normative ideas that infuse our encounters with deepfakes. Within this, I further raise the question of what moral ground deepfakes occupy that allows them to have such a potentially devastating effect. I show that answering this question involves reinstating the notion that photographs are popularly conceived of as transparent. The rejoinder to this argument, however, is that to take the sting out of deepfakes we must, once again, become skeptical of the veracity of all images, including photoreal ones. This kind of critical mindedness was warranted before the invention of photography for both pictorial imprints and written accounts from various media sources. Given this, along with the fact that photographic trickery is nothing new, deepfakes need not push us into a post-truth epistemic abyss, for they imply a decidedly old turn. (shrink)

Deep fakes, a technology enabling the creation of highly realistic fake images and videos through face-swapping, have sparked significant attention due to their potential for misuse. This paper explores the technologies behind deep fakes and categorizes them into four types: deep fake pornography, political campaigns, commercial uses, and creative content. The authors discuss the ethical and regulatory challenges associated with each category. Initially, deep fakes were used maliciously, such as in revenge porn and political manipulation, which led to widespread fear. (...) However, like many technologies before it, deep fakes may eventually become more widely accepted. The paper also examines the role of online platforms and governments in addressing these issues and offers an optimistic view of how the democratization of AI might help resolve the ethical concerns surrounding deep fakes. -/- . (shrink)

Deep fakes, a technology enabling the creation of highly realistic fake images and videos through face-swapping, have sparked significant attention due to their potential for misuse. This paper explores the technologies behind deep fakes and categorizes them into four types: deep fake pornography, political campaigns, commercial uses, and creative content. The authors discuss the ethical and regulatory challenges associated with each category. Initially, deep fakes were used maliciously, such as in revenge porn and political manipulation, which led to widespread fear. (...) However, like many technologies before it, deep fakes may eventually become more widely accepted. The paper also examines the role of online platforms and governments in addressing these issues and offers an optimistic view of how the democratization of AI might help resolve the ethical concerns surrounding deep fakes. (shrink)

Deepfakes are audio, video, or still-image digital artifacts created by the use of artificial intelligence technology, as opposed to traditional means of recording. Because deepfakes can look and sound much like genuine digital recordings, they have entered the popular imagination as sources of serious epistemic problems for us, as we attempt to navigate the increasingly treacherous digital information environment of the internet. In this paper, I attempt to clarify what epistemic problems deepfakes pose and why they pose these problems, by (...) drawing parallels between recordings and our own senses as sources of evidence. I show that deepfakes threaten to undermine the status of digital recordings as evidence. The existence of deepfakes thus encourages a kind of skepticism about digital recordings that bears important similarities to classic philosophical skepticism concerning the senses. However, the skepticism concerning digital recordings that deepfakes motivate is also importantly different from classical skepticism concerning the senses, and I argue that these differences illuminate some possible strategies for solving the epistemic problems posed by deepfakes. (shrink)

Synthetic media generators, such as DALL-E, and synthetic media artifacts, such as deepfakes, undermine our fundamental epistemic standards and practices. Yet, the nature of their epistemic threat remains elusive. After all, fictional or distorted representations of reality are as old as photography. We argue that the novel epistemic threat of synthetic media is that, for the first time, synthetic media tools afford ordinary computer users the practicable possibility to cheaply and effortlessly create and widely share fictional worlds indistinguishable from the (...) real world or credible representations of it. We further argue that a synthetic media artifact is epistemically malignant in a given media context for a person acquainted with the context when the person is misled to confuse the version of the world depicted in it with the real world in an epistemically or morally significant way. (shrink)

דמיון מתחולל בחלקו מחוץ לראש באמצעות עזרים טכנולוגיים. עובדה זו כשלעצמה אינה חדשה, אולם ‏טכנולוגיות חדשות משנות את טיבו של הדמיון. מחוללי מדיה סינתטית, כגון דָאלִי, ופריטי מדיה סינתטיים, ‏כגון זיופים עמוקים (‏דיפ-פייקס; תמונות וסרטונים ריאליסטיים שנוצרו באופן אלגוריתמי ושמציגים אנשים ‏מבצעים או אומרים משהו שלא עשו או אמרו) מערערים על אמות המידה האפיסטמיות הבסיסיות שלנו. עם ‏זאת, טבעו של האיום האפיסטמי החדש הנשקף מהם נותר חמקמק, שכן ייצוגים בדיוניים או מעוותים של ‏המציאות הם עתיקים לפחות כמו הצילום עצמו. אפיונים (...) קיימים באפיסטמולוגיה חברתית של האיום הגלום ‏בזיופים עמוקים מתקשים לזהות מה בעצם חדש בו. למאמר זה שתי מטרות: לאפיין את האיום האפיסטמי ‏החדש הנשקף ממדיה סינתטית, ולזהות את התנאים שבהם פריט מדיה סינתטית הוא בעייתי מבחינה ‏אפיסטמית. אנו טוענים שהאיום האפיסטמי הגלום בזיופים עמוקים בפרט, ובמדיה סינתטית בכלל, הוא ‏שלראשונה למשתמשי מחשב רגילים זמינה האפשרות המעשית ליצור בזול וללא מאמץ עולמות בדיוניים שלא ‏ניתן להבחין ביניהם לבין העולם האמיתי ולשתפם בתפוצה רחבה. פריט מדיה סינתטי הוא רעיל מבחינה ‏אפיסטמית בהקשר נתון כאשר אדם בעל הכרה כלשהי של ההקשר עלול לבלבל את העולם הבדיוני המתואר בו ‏עם העולם הממשי או עם ייצוגים מהימנים שלו בנושאים בעלי חשיבות מוסרית או אפיסטמית. ‏ -/- Imagination takes place partly outside the mind, through technological aids. This fact is not new in ‎itself, but new technologies are changing the nature of imagination. Synthetic media generators, ‎such as DALL·E, and synthetic media items, such as deepfakes (realistic images and videos ‎algorithmically generated to depict people doing or saying things they never actually did or said), ‎challenge our fundamental epistemic standards. But the nature of the new epistemic threat they ‎pose remains elusive, since fictional or distorted representations of reality are at least as old as ‎photography itself. Existing characterizations in social epistemology of the threat posed by ‎deepfakes struggle to pinpoint what exactly is novel about it. This paper has two aims: to ‎characterize the new epistemic threat posed by synthetic media, and to identify the conditions ‎under which a synthetic media item is epistemically malignant or toxic. We argue that the ‎epistemic threat posed by deepfakes in particular, and by synthetic media more broadly, lies in the ‎unprecedented practicable possibility afforded to ordinary computer users to cheaply and ‎effortlessly create fictional worlds that are indistinguishable from the real world and share them ‎widely. A synthetic media item is epistemically toxic or malignant in a given context when a person ‎acquainted with the context is liable to confuse the fictional world depicted in it with the real ‎world or with credible representations thereof on matters of moral or epistemic significance.‎. (shrink)

A novel deep learning architecture for deepfake detection was proposed that combines LSTM and CNN methods, implemented using Python on the Kaggle platform. The model was evaluated using two datasets: DFDC and Ciplab. Dataset preprocessing involved 19,148 real images and an equal number of fake images, with 80% allocated for training using 128 × 128 image sizes. Binary cross-entropy function 5.4 was used to calculate error rates during training iterations. The results demonstrated high accuracy rates of 97.32% and 98.24%, (...) with low error rates of 0.15% and 0.26% for the respective datasets. These results confirm the model's adaptability for accurate deepfake prediction. The effectiveness stems from the complementary strengths of both approaches: LSTMs effectively capture sequential temporal dependencies between frames to identify deepfake inconsistencies, while CNNs evaluate frame content to detect visual artifacts, inconsistencies, or unusual patterns indicating manipulation. The hybrid approach leverages the strengths of both methods, resulting in a more efficient and effective model capable of adapting to various deepfake generation techniques. By combining temporal analysis capabilities of LSTMs with the spatial feature extraction abilities of CNNs, the architecture achieves a more comprehensive deepfake detection system. (shrink)

The Orpheus Hypothesis I (Graziano, 2026) established rhythm as the ontological ground of musical experience, while the T.I.C.A. model formalized the indeterminacy of temporal coupling between dynamical systems. These findings, taken together, generate a consequence that their authors did not anticipate — or perhaps anticipated too clearly to state without preparation. They collide, with considerable force, against a fundamental historiographical anomaly: the alleged teleological acceleration of Western music. The history of a phenomenon that is, at its ontological core, an embodied, (...) indeterminate, non-linear event cannot itself be a linear, necessary, and progressive sequence. Something is wrong with the story we have been told. This paper proposes that the 'Western Canon' is not an organic evolutionary process but a twentieth-century simulation — a project of Temporal Engineering — and that the history of music was synchronized at the dawn of the phonographic era to facilitate mass temporal entrainment on an unprecedented scale. If the Orpheus Hypothesis is correct, music history is a retroactive deepfake designed to replace biological resonance with industrial synchronization. The paper proceeds in five movements, each of which removes one load-bearing pillar from the edifice of musical historiography, and concludes by proposing the Orpheus Hypothesis not merely as a theoretical framework but as an instrument of diagnosis — and, for those who choose to use it as such, of liberation. (shrink)

The rapid advancement of generative artificial intelligence (AI) has introduced unprecedented challenges in distinguishing authentic human faces from synthetically generated counterparts. Existing research predominantly focuses on binary classification — real versus fake — without differentiating between distinct forgery mechanisms. This paper presents the first publicly available multi-class face forgery dataset comprising four categories: real photographs (1,081 images), Photoshop-manipulated faces (960 images), Generative Adversarial Network (GAN)-generated faces (2,000 images), and Diffusion model-generated faces (2,000 images), totaling 6,041 images. We conduct a comprehensive (...) comparative evaluation of three state-of-the-art deep learning architectures — ResNet-50, EfficientNet-B4, and Swin Transformer (Swin-T) — under identical experimental conditions. The Swin Transformer achieves the highest test accuracy of 93.38%, with a macro-average F1-score of 90.32%, demonstrating superior capability in capturing fine-grained forgery artifacts. The results underscore that the Swin Transformer’s superiority stems from its ability to model long-range spatial dependencies, which proved instrumental in detecting global artifacts characteristic of diffusion- based synthesis. Notably, GAN and Diffusion-generated faces are detected with 100% accuracy, while Photoshop-manipulated faces remain the most challenging category (F1: 80%). Gradient-weighted Class Activation Mapping (Grad-CAM) visualizations reveal distinct spatial attention patterns per forgery type, providing interpretable insights into model decision-making. The dataset is publicly released on Kaggle to facilitate future research in multi-class deepfake detection. (shrink)

The interpretation of many texts in the everyday world is concerned with their truth value in relation to the reality around us. The recent publication experiments with computer-generated texts have shown that the distinction between true and false, or reality and fiction, is not always clear from the text itself. Essentially, in today’s media space, one may encounter texts, videos or images that deceive the reader by displaying nonsensical content or nonexistent events, while nevertheless appearing as genuine human-produced messages. This (...) article outlines certain problems with artificial intelligence (AI)-generated content, and frames the issue as a problem of recognising its proper referential reality. Examples include generative texts, deepfakes and their functioning in contemporary culture. The article makes use of the concepts of mimicry and nonsense to reveal the elements and counterparts in the communicative processes involving generated media. (shrink)

This paper examines the transformative shifts in journalism driven by the widespread integration of GenAI in modern media. The digital convergence requires new effective solutions for the verification process of facts, images and especially videos. The rise of sophisticated deepfake videos complicates the gatekeeping role of journalists. Increasingly, first-hand information and the personalization of both the journalist and the media outlet remain pivotal for audience trust. By analyzing current limitations of existing detection tools for deepfakes, this study explores how (...) newsrooms should adapt their workflows to preserve public trust. This paper explains how deepfake videos could be exposed through a combination of metadata and visual clues. It offers a practical framework for newsrooms to implement tools for technical proof. (shrink)

Generative Artificial Intelligence (AI) has revolutionized numerous industries by enabling the creation of realistic and novel content across various mediums, including art, music, writing, and video. However, with this power comes a significant set of ethical challenges. These challenges revolve around issues such as authorship, deception, bias, and the potential for misuse. This paper explores the ethical concerns surrounding generative AI, examining the delicate balance between creativity and deception. By analyzing the implications of AI-generated content, including deepfakes and synthetic media, (...) the study investigates the risks associated with AI in terms of misinformation, intellectual property rights, and the authenticity of creative work. Additionally, the paper evaluates current ethical frameworks and proposes guidelines for the responsible use of generative AI. Ultimately, it aims to provide a comprehensive overview of the ethical landscape in the realm of AI-generated content. (shrink)

Deepfakes and other forms of synthetic media are widely regarded as serious threats to our knowledge of the world. Various technological responses to these threats have been proposed. The reactive approach proposes to use artificial intelligence to identify synthetic media. The proactive approach proposes to use blockchain and related technologies to create immutable records of verified media content. I argue that both approaches, but especially the reactive approach, are vulnerable to a problem analogous to the ancient problem of the criterion—a (...) line of argument with skeptical implications. I argue that, while the proactive approach is relatively resistant to this objection, it faces its own serious challenges. In short, the proactive approach would place a heavy burden on users to verify their own content, a burden that is exacerbated by and is likely to exacerbate existing inequalities. (shrink)

In the presently unfolding deepfake era, recurrent inflationary algorithmic superintelligence (ASI) achievement claims degenerated from being a mere reflection of an exaggerated but candid initial enthusiasm to becoming a convenient tool for misdirection facilitating epistemic perpetuum mobile (EPM) scams. This transdisciplinarily conceived paper compactly analyzes the underlying ASI definition avoidance problem which emerged from interactions between three major epistemic trends in the ASI debate: boomerism, doomerism and pragmatism. Via taking a fourth external perspective entertained by a fictive entity called (...) Cyogenes who volunteers as a philosopher and criticizes contemporary conventions (inspired by the ancient Cynicism lived by early Greek philosophers including Antisthenes and Diogenes of Sinope), the paper highlights key strengths but also particularly pitfalls within those three aforementioned epistemic trends all of which face the ASI definition avoidance problem – clearing the way for global EPM scams. After this philosophical analysis, a deflationary account of ASI (i.e. the epistemic perpetuum mobile) is discussed. Thereafter, practical cybersecurity-oriented threat models and defenses against EPM scams are formulated. In brief, fruitless EPM scams unnecessarily risk intensifying the near-term existential risk of epistemic self-sabotage and reinforcing totalitarian aspirations. (shrink)

In the presently unfolding deepfake era, previously unrelated algorithmic superintelligence possibility claims cannot be scientifically analyzed in isolation anymore due to the connected inevitable epistemic interactions that have already commenced. For instance, deep-learning (DL) related algorithmic supremacy claims may intrinsically compete with both neuro-symbolic (NS) algorithmic and further quantum (Q) algorithmic superintelligence achievement claims. Concurrently, a variety of experimental combinations of DL, NS and Q directions are conceivable. While research on these three illustrative variants did not yet offer any (...) clear mutually applicable scientific definition of a purported supremacy level framed as goal, the currently most robust scientific evaluation frameworks would i.a. require a joint testing against pre-existing non-algorithmic supremacy baselines such as e.g. conscious supremacy and cellular supremacy. This autodidactic paper explains why, for scientific reasons, all algorithmic superintelligence achievement claims in the deepfake era can be condensed and subjected to a joint scientific evaluation framework comprising multiple successive civilization-level tasks of epistemic tunneling (ET). In light of this elusive requirement, recent claims of science automation are spurious and misleading. It is vital not to misapprehend algorithmic epistemic dark matter (EDM) mining and epistemic dark energy (EDE) generation as paradigm-shifting ET events. (shrink)

In the deepfake era, materialism and idealism seem to clash at multiple epistemic levels with new additional facets unfolding – an epistemic friction which could act as creativity-stimulating impetus for science and philosophy. Could the information-related concept of supercomplexity be instrumental in better clarifying understudied aspects of the apparent dichotomy? Instead of directly answering this question, this short autodidactic paper compactly analyzes a small but potentially relevant puzzle piece to complexity research taking the form of an explanatory bridge from (...) complexity notions inherently focused on lifeless matter to a multi-dimensional informational supercomplexity concept most facets of which have hitherto been neglected. When considering the conjunction of biocosmological and cyborgnetic arguments, it seems that life, consciousness and the practice of science itself need not be a priori excluded from a physically-grounded scientific complexity theory. (shrink)

In the 21st century, Chinese universities stand at a historic crossroads. -/- On one hand, they are striving mightily—proclaiming the ambition to build “world-class universities,” pouring massive investments into artificial intelligence, quantum technology, aerospace engineering, and life sciences, aggressively recruiting international talent and launching global partnerships. This is their DeepSeek: the profound pursuit of global knowledge and scientific peaks. -/- On the other hand, they are increasingly trapped in systemic falsification—data fabrication, academic misconduct, loss of academic freedom, ideological rigidity, hollow (...) internationalization, and performative pedagogy have become commonplace. This is their DeepFake: a simulation of a “real university,” carefully manufactured and strategically displayed. -/- This paper argues: The dual existence of DeepSeek and DeepFake is no coincidence. They are twin products of the same political and cultural machinery. When the university becomes an instrument of political performance and national image-building, the pursuit of truth inevitably degenerates into a spectacle of appearances. (shrink)

While the present information ecosystem is still undergoing a tsunami of repeated algorithmic superintelligence (ASI) achievement claims linked to the motif of the epistemic perpetuum mobile (EPM), will the laterally emerging and slowly propagating quantum ASI hype finally lead to a multiversal fear of missing out? Instead of adding novel entries to the already large enough and growing set of prophecies about the future promulgated in the deepfake era, this paper written for purposes of self-education utilizes a recent epistemic (...) complexity theory to answer the following natural scientific question: could the impossibility to build a multiversal ASI be amenable to experimental problematization? (shrink)

Amidst the flow of repetitive inflationary algorithmic superintelligence (ASI) achievement claims linked to the rise of epistemic perpetuum mobile (EPM) scams in the deepfake era, this compact transdisciplinary paper merely written as ephemeral mental clipboard for the purpose of self-education collates the results of a new epistemic framework to achieve a better clarification of the present widespread confusion.