Dataset Information

Fooled twice: People cannot detect deepfakes but think they can.

ABSTRACT: Hyper-realistic manipulations of audio-visual content, i.e., deepfakes, present new challenges for establishing the veracity of online content. Research on the human impact of deepfakes remains sparse. In a pre-registered behavioral experiment (N = 210), we show that (1) people cannot reliably detect deepfakes and (2) neither raising awareness nor introducing financial incentives improves their detection accuracy. Zeroing in on the underlying cognitive processes, we find that (3) people are biased toward mistaking deepfakes as authentic videos (rather than vice versa) and (4) they overestimate their own detection abilities. Together, these results suggest that people adopt a "seeing-is-believing" heuristic for deepfake detection while being overconfident in their (low) detection abilities. The combination renders people particularly susceptible to be influenced by deepfake content.

SUBMITTER: Kobis NC

PROVIDER: S-EPMC8602050 | biostudies-literature | 2021 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Fooled twice: People cannot detect deepfakes but think they can.

Köbis Nils C NC Doležalová Barbora B Soraperra Ivan I

iScience 20211029 11

Hyper-realistic manipulations of audio-visual content, i.e., deepfakes, present new challenges for establishing the veracity of online content. Research on the human impact of deepfakes remains sparse. In a pre-registered behavioral experiment (<i>N</i> = 210), we show that (1) people cannot reliably detect deepfakes and (2) neither raising awareness nor introducing financial incentives improves their detection accuracy. Zeroing in on the underlying cognitive processes, we find that (3) people a ...[more]

PMID: 34820608

Similar Datasets

Project description:The recent SARS-CoV-2 epidemic has highlighted the need to prevent emerging and re-emerging diseases, which means that we must approach the study of diseases from a One Health perspective. The study of pathogen transmission in wildlife is challenging, but it is unquestionably key to understand how epidemiological interactions occur at the wildlife-domestic-human interface. In this context, studying parasite avoidance behaviours may provide essential insights on parasite transmission, host-parasite coevolution, and energy flow through food-webs. However, the strategies of avoiding trophically transmitted parasites in mammalian carnivores have received little scientific attention. Here, we explore the behaviour of red foxes (Vulpes vulpes) and other mammalian carnivores at conspecific and heterospecific carnivore carcasses using videos recorded by camera traps. We aim to determine 1) the factors influencing the probability of foxes to practice cannibalism, and 2) whether the scavenging behaviour of foxes differ when facing conspecific vs. heterospecific carcasses. We found that red foxes were generally reluctant to consume mesocarnivore carrion, especially of conspecifics. When recorded, consumption by foxes was delayed several days (heterospecific carcasses) or weeks (conspecific carcasses) after carcass detection. Other mammalian scavengers showed a similar pattern. Also, meat-borne parasite transmission from wild carnivore carcasses to domestic dogs and cats was highly unlikely. Our findings challenge the widespread assumption that cannibalistic or intra-specific scavenging is a major transmission route for Trichinella spp. and other meat-borne parasites, especially for the red fox. Overall, our results suggest that the feeding decisions of scavengers are probably shaped by two main contrasting forces, namely the nutritional reward provided by carrion of phylogenetically similar species and the risk of acquiring meat-borne parasites shared with these species. This study illustrates how the detailed monitoring of carnivore behaviour is essential to assess the epidemiological role of these hosts in the maintenance and dispersion of parasites of public and animal health relevance.

Project description:A deepfake is content or material that is synthetically generated or manipulated using artificial intelligence (AI) methods, to be passed off as real and can include audio, video, image, and text synthesis. The key difference between manual editing and deepfakes is that deepfakes are AI generated or AI manipulated and closely resemble authentic artifacts. In some cases, deepfakes can be fabricated using AI-generated content in its entirety. Deepfakes have started to have a major impact on society with more generation mechanisms emerging everyday. This article makes a contribution in understanding the landscape of deepfakes, and their detection and generation methods. We evaluate various categories of deepfakes especially in audio. The purpose of this survey is to provide readers with a deeper understanding of (1) different deepfake categories; (2) how they could be created and detected; (3) more specifically, how audio deepfakes are created and detected in more detail, which is the main focus of this paper. We found that generative adversarial networks (GANs), convolutional neural networks (CNNs), and deep neural networks (DNNs) are common ways of creating and detecting deepfakes. In our evaluation of over 150 methods, we found that the majority of the focus is on video deepfakes, and, in particular, the generation of video deepfakes. We found that for text deepfakes, there are more generation methods but very few robust methods for detection, including fake news detection, which has become a controversial area of research because of the potential heavy overlaps with human generation of fake content. Our study reveals a clear need to research audio deepfakes and particularly detection of audio deepfakes. This survey has been conducted with a different perspective, compared to existing survey papers that mostly focus on just video and image deepfakes. This survey mainly focuses on audio deepfakes that are overlooked in most of the existing surveys. This article's most important contribution is to critically analyze and provide a unique source of audio deepfake research, mostly ranging from 2016 to 2021. To the best of our knowledge, this is the first survey focusing on audio deepfakes generation and detection in English.

Dataset Information

Fooled twice: People cannot detect deepfakes but think they can.

Publications

Fooled twice: People cannot detect deepfakes but think they can.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets