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)

Deepfakes are fake recordings generated by machine learning algorithms. Various philosophical explanations have been proposed to account for their epistemic harmfulness. In this paper, I argue that deepfakes are epistemically harmful because they undermine trust in recording technology. As a result, we are no longer entitled to our default doxastic attitude of believing that P on the basis of a recording that supports the truth of P. Distrust engendered by deepfakes changes the epistemic status of recordings to resemble that of (...) handmade images. Their credibility, like that of testimony, depends partly on the credibility of the source. I consider some proposed technical solutions from a philosophical perspective to show the practical relevance of these suggestions. (shrink)

Recent technological innovation has made video doctoring increasingly accessible. This has given rise to Deepfake Pornography, an emerging phenomenon in which Deep Learning algorithms are used to superimpose a person’s face onto a pornographic video. Although to most people, Deepfake Pornography is intuitively unethical, it seems difficult to justify this intuition without simultaneously condemning other actions that we do not ordinarily find morally objectionable, such as sexual fantasies. In the present article, I refer to this contradiction as the pervert’s dilemma. (...) I propose that the method of Levels of Abstraction, a philosophical mode of enquiry inspired by Formal Methods in computer science, can be employed to formulate at least one possible solution to the dilemma. From this perspective, the permissibility of some actions appears to depend on the degree to which they are abstracted from their natural context. I conclude that the dilemma can only be solved when considered at low levels of abstractions, when Deepfakes are situated in the macro-context of gender inequality. (shrink)

This paper defends two main theses related to emerging deepfake technology. First, fears that deepfakes will bring about epistemic catastrophe are overblown. Such concerns underappreciate that the evidential power of video derives not solely from its content, but also from its source. An audience may find even the most realistic video evidence unconvincing when it is delivered by a dubious source. At the same time, an audience may find even weak video evidence compelling so long as it is delivered by (...) a trusted source. The growing prominence of deepfake content is unlikely to change this fundamental dynamic. Thus, through appropriate patterns of trust, whatever epistemic threat deepfakes pose can be substantially mitigated. The second main thesis is that focusing on deepfakes that are intended to deceive, as epistemological discussions of the technology tend to do, threatens to overlook a further effect of this technology. Even where deepfake content is not regarded by its audience as veridical, it may cause its viewers to develop psychological associations based on that content. These associations, even without rising to the level of belief, may be harmful to the individuals depicted and more generally. Moreover, these associations may develop in cases in which the video content is realistic, but the audience is dubious of the content in virtue of skepticism toward its source. Thus, even if—as I suggest—epistemological concerns about deepfakes are overblown, deepfakes may nonetheless be psychologically impactful and may do great harm. (shrink)

Many of our beliefs are acquired online. Online epistemic environments are replete with fake news, fake science, fake photographs and videos, and fake people in the form of trolls and social bots. The purpose of this paper is to investigate the threat that such online fakes pose to the acquisition of knowledge. I argue that fakes can interfere with one or more of the truth, belief, and warrant conditions on knowledge. I devote most of my attention to the effects of (...) online fakes on satisfaction of the warrant condition, as these have received comparatively little attention. I consider three accounts of the conditions under which fakes compromise the warrant condition. I argue for the third of these accounts, according to which the propensity of fakes to exist in an environment threatens warrant acquisition in that environment. Finally, I consider some limitations on the epistemic threat of fakes and suggest some strategies by which this threat can be mitigated. (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)

[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)

I shall first briefly revisit the broad idea of ‘epistemic injustice’, explaining how it can take either distributive or discriminatory form, in order to put the concepts of ‘testimonial injustice’ and ‘hermeneutical injustice’ in place. In previous work I have explored how the wrong of both kinds of epistemic injustice has both an ethical and an epistemic significance—someone is wronged in their capacity as a knower. But my present aim is to show that this wrong can also have a political (...) significance in relation to non-domination, and so to freedom. While it is only the republican conception of political freedom that presents nondomination as constitutive of freedom, I shall argue that non-domination is best understood as a thoroughly generic liberal ideal of freedom to which even negative libertarians are implicitly committed, for non-domination is negative liberty as of right—secured non-interference. Crucially on this conception, non-domination requires that the citizen can contest interferences. Pettit specifies three conditions of contestation, each of which protects against a salient risk of the would-be contester not getting a ‘proper hearing’. But I shall argue that missing from this list is anything to protect against a fourth salient threat: the threat that either kind of epistemic injustice might disable contestation by way of an unjust deflation of either credibility or intelligibility. Thus we see that both testimonial and hermeneutical injustice can render a would-be contester dominated. Epistemic justice is thereby revealed as a constitutive condition of non-domination, and thus of a central liberal political ideal of freedom. (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)

Deepfakes are realistic videos created using new machine learning techniques rather than traditional photographic means. They tend to depict people saying and doing things that they did not actually say or do. In the news media and the blogosphere, the worry has been raised that, as a result of deepfakes, we are heading toward an “infopocalypse” where we cannot tell what is real from what is not. Several philosophers have now issued similar warnings. In this paper, I offer an analysis (...) of why deepfakes are such a serious threat to knowledge. Utilizing the account of information carrying recently developed by Brian Skyrms, I argue that deepfakes reduce the amount of information that videos carry to viewers. I conclude by drawing some implications of this analysis for addressing the epistemic threat of deepfakes. (shrink)

Schlick held that our knowledge is founded upon certain contingent apprehensions which he described as follows: "I grasp their meaning at the same time that I grasp their truth." He cites as an example the apprehension expressed by "Yellow here now." When such apprehensions are expressed in syntactically well-formed sentences, they can be seen to have certain psychological states as their objects - and therefore to be similar in all essential respects to what members of the Brentano school had called (...) "inner perceptions.". (shrink)

One of the many problems that would have t o be solved by a satisfactory theory of empirical knowledge, perhaps the most central is a general structural problem which I shall call the epistemic regress problem: the problem of how to avoid an in- finite and presumably vicious regress of justification in ones account of the justifica- tion of empirical beliefs. Foundationalist theories of empirical knowledge, as we shall see further below, attempt t o avoid the regress by locating a (...) class of empirical beliefs whose justification does not depend on that of other empirical beliefs. Extemalist theories, the topic of the present paper, represent one species of foundationalism. (shrink)

In Epistemic injustice, Miranda Fricker makes a tremendous contribution to theorizing the intersection of social epistemology with theories of justice. Theories of justice often take as their object of assessment either interpersonal transactions (specific exchanges between persons) or particular institutions. They may also take a more comprehensive perspective in assessing systems of institutions. This systemic perspective may enable control of the cumulative effects of millions of individual transactions that cannot be controlled at the individual or institutional levels. This is true (...) not only with respect to the overall distribution of such goods as income and wealth, but also with respect to the goods of testimonial and hermeneutical justice. Cognitive biases that may be difficult for even epistemically virtuous individuals to correct on their own may be more susceptible to correction if we focus on the principles that should govern our systems of testimonial gathering and assessment. Hence, while Fricker’s focus on individual epistemic virtue is important, we also need to consider what epistemic justice as a virtue of social systems would require. My paper will indicate some directions forward on this front, focusing on the need for integration of diverse institutions and persons engaged in inquiry. (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)

What modal relation must a fact bear to a belief in order for this belief to constitute knowledge of that fact? Externalists have proposed various answers, including some that combine externalism with contextualism. We shall find that various forms of externalism share a modal conception of “sensitivity” open to serious objections. Fortunately, the undeniable intuitive attractiveness of this conception can be explained through an easily confused but far preferable notion of “safety.” The denouement of our reflections, finally, will be to (...) show how replacing sensitivity with safety makes it possible to defend plain Moorean common sense against the spurious advantages over it claimed by skeptical, tracking, relevant-alternative, and contextualist accounts. (shrink)