Tampered multimedia content is being increasingly used in a broad range of cybercrime activities. The spread of fake news, misinformation, digital kidnapping, and ransomware-related crimes are amongst the most recurrent crimes in which manipulated digital photos and videos are the perpetrating and disseminating medium. Criminal investigation has been challenged in applying machine learning techniques to automatically distinguish between fake and genuine seized photos and videos. Despite the pertinent need for manual validation, easy-to-use platforms for digital forensics are essential to automate and facilitate the detection of tampered content and to help criminal investigators with their work. This paper presents a machine learning Support Vector Machines (SVM) based method to distinguish between genuine and fake multimedia files, namely digital photos and videos, which may indicate the presence of deepfake content. The method was implemented in Python and integrated as new modules in the widely used digital forensics application Autopsy. The implemented approach extracts a set of simple features resulting from the application of a Discrete Fourier Transform (DFT) to digital photos and video frames. The model was evaluated with a large dataset of classified multimedia files containing both legitimate and fake photos and frames extracted from videos. Regarding deepfake detection in videos, the Celeb-DFv1 dataset was used, featuring 590 original videos collected from YouTube, and covering different subjects. The results obtained with the 5-fold cross-validation outperformed those SVM-based methods documented in the literature, by achieving an average F1-score of 99.53%, 79.55%, and 89.10%, respectively for photos, videos, and a mixture of both types of content. A benchmark with state-of-the-art methods was also done, by comparing the proposed SVM method with deep learning approaches, namely Convolutional Neural Networks (CNN). Despite CNN having outperformed the proposed DFT-SVM compound method, the competitiveness of the results attained by DFT-SVM and the substantially reduced processing time make it appropriate to be implemented and embedded into Autopsy modules, by predicting the level of fakeness calculated for each analyzed multimedia file.

BACKGROUND: The digital era has witnessed an escalating dependence on digital platforms for news and information, coupled with the advent of \"deepfake\" technology. Deepfakes, leveraging deep learning models on extensive data sets of voice recordings and images, pose substantial threats to media authenticity, potentially leading to unethical misuse such as impersonation and the dissemination of false information.
OBJECTIVE: To counteract this challenge, this study aims to introduce the concept of innate biological processes to discern between authentic human voices and cloned voices. We propose that the presence or absence of certain perceptual features, such as pauses in speech, can effectively distinguish between cloned and authentic audio.
METHODS: A total of 49 adult participants representing diverse ethnic backgrounds and accents were recruited. Each participant contributed voice samples for the training of up to 3 distinct voice cloning text-to-speech models and 3 control paragraphs. Subsequently, the cloning models generated synthetic versions of the control paragraphs, resulting in a data set consisting of up to 9 cloned audio samples and 3 control samples per participant. We analyzed the speech pauses caused by biological actions such as respiration, swallowing, and cognitive processes. Five audio features corresponding to speech pause profiles were calculated. Differences between authentic and cloned audio for these features were assessed, and 5 classical machine learning algorithms were implemented using these features to create a prediction model. The generalization capability of the optimal model was evaluated through testing on unseen data, incorporating a model-naive generator, a model-naive paragraph, and model-naive participants.
RESULTS: Cloned audio exhibited significantly increased time between pauses (P<.001), decreased variation in speech segment length (P=.003), increased overall proportion of time speaking (P=.04), and decreased rates of micro- and macropauses in speech (both P=.01). Five machine learning models were implemented using these features, with the AdaBoost model demonstrating the highest performance, achieving a 5-fold cross-validation balanced accuracy of 0.81 (SD 0.05). Other models included support vector machine (balanced accuracy 0.79, SD 0.03), random forest (balanced accuracy 0.78, SD 0.04), logistic regression, and decision tree (balanced accuracies 0.76, SD 0.10 and 0.72, SD 0.06). When evaluating the optimal AdaBoost model, it achieved an overall test accuracy of 0.79 when predicting unseen data.
CONCLUSIONS: The incorporation of perceptual, biological features into machine learning models demonstrates promising results in distinguishing between authentic human voices and cloned audio.

The rapid advancement of deepfake technology poses an escalating threat of misinformation and fraud enabled by manipulated media. Despite the risks, a comprehensive understanding of deepfake detection techniques has not materialized. This research tackles this knowledge gap by providing an up-to-date systematic survey of the digital forensic methods used to detect deepfakes. A rigorous methodology is followed, consolidating findings from recent publications on deepfake detection innovation. Prevalent datasets that underpin new techniques are analyzed. The effectiveness and limitations of established and emerging detection approaches across modalities including image, video, text and audio are evaluated. Insights into real-world performance are shared through case studies of high-profile deepfake incidents. Current research limitations around aspects like cross-modality detection are highlighted to inform future work. This timely survey furnishes researchers, practitioners and policymakers with a holistic overview of the state-of-the-art in deepfake detection. It concludes that continuous innovation is imperative to counter the rapidly evolving technological landscape enabling deepfakes.

Video recordings accurately capture facial expression movements; however, they are difficult for face perception researchers to standardise and manipulate. For this reason, dynamic morphs of photographs are often used, despite their lack of naturalistic facial motion. This study aimed to investigate how humans perceive emotions from faces using real videos and two different approaches to artificially generating dynamic expressions - dynamic morphs, and AI-synthesised deepfakes. Our participants perceived dynamic morphed expressions as less intense when compared with videos (all emotions) and deepfakes (fearful, happy, sad). Videos and deepfakes were perceived similarly. Additionally, they perceived morphed happiness and sadness, but not morphed anger or fear, as less genuine than other formats. Our findings support previous research indicating that social responses to morphed emotions are not representative of those to video recordings. The findings also suggest that deepfakes may offer a more suitable standardized stimulus type compared to morphs. Additionally, qualitative data were collected from participants and analysed using ChatGPT, a large language model. ChatGPT successfully identified themes in the data consistent with those identified by an independent human researcher. According to this analysis, our participants perceived dynamic morphs as less natural compared with videos and deepfakes. That participants perceived deepfakes and videos similarly suggests that deepfakes effectively replicate natural facial movements, making them a promising alternative for face perception research. The study contributes to the growing body of research exploring the usefulness of generative artificial intelligence for advancing the study of human perception.

Recently, there have been notable advancements in video editing software. These advancements have allowed novices or those without access to advanced computer technology to generate videos that are visually indistinguishable to the human eye from real ones to the human observer. Therefore, the application of deepfake technology has the potential to expand the scope of identity theft, which poses a significant risk and a formidable challenge to global security. The development of an effective approach for detecting fake videos is necessary. Here, we introduce a novel methodology that employs a convolutional neural network (CNN) and Gaussian mixture model (GMM) to effectively differentiate between fake and real images or videos. The proposed methodology presents a novel CNN-GMM architecture in which the fully connected (FC) layer in the CNN is replaced with a customized Gaussian mixture model (GMM) fully connected layer. The GMM layer utilizes a weighted set of Gaussian probability density functions (PDFs) to represent the distribution of data frequencies in both real and fake images. This representation indicates there is a shift in the distribution of the manipulated images due to added noise. The CNN-GMM model demonstrates the ability to accurately identify variations resulting from different types of deepfakes within the probability distribution. It achieves a high level of classification accuracy, reaching up to 100% in training accuracy and up to 96% in validation accuracy. Notwithstanding the ratio of the genuine class to the counterfeit class being 16.6% to 83.4%, the CNN-GMM model exhibited high-performance metrics in terms of recall, accuracy, and F-score when classifying the least genuine class.

In recent discussions in the European Parliament, the need for regulations for so-called high-risk artificial intelligence (AI) systems was identified, which are currently codified in the upcoming EU Artificial Intelligence Act (AIA) and approved by the European Parliament. The AIA is the first document to be turned into European Law. This initiative focuses on turning AI systems in decision support systems (human-in-the-loop and human-in-command), where the human operator remains in control of the system. While this supposedly solves accountability issues, it includes, on one hand, the necessary human-computer interaction as a potential new source of errors; on the other hand, it is potentially a very effective approach for decision interpretation and verification. This paper discusses the necessary requirements for high-risk AI systems once the AIA comes into force. Particular attention is paid to the opportunities and limitations that result from the decision support system and increasing the explainability of the system. This is illustrated using the example of the media forensic task of DeepFake detection.

The rapid advancement of \'deepfake\' video technology-which uses deep learning artificial intelligence algorithms to create fake videos that look real-has given urgency to the question of how policymakers and technology companies should moderate inauthentic content. We conduct an experiment to measure people\'s alertness to and ability to detect a high-quality deepfake among a set of videos. First, we find that in a natural setting with no content warnings, individuals who are exposed to a deepfake video of neutral content are no more likely to detect anything out of the ordinary (32.9%) compared to a control group who viewed only authentic videos (34.1%). Second, we find that when individuals are given a warning that at least one video in a set of five is a deepfake, only 21.6% of respondents correctly identify the deepfake as the only inauthentic video, while the remainder erroneously select at least one genuine video as a deepfake.

This study is one of the first to investigate the relationship between modalities and individuals\' tendencies to believe and share different forms of deepfakes (also deep fakes). Using an online survey experiment conducted in the US, participants were randomly assigned to one of three disinformation conditions: video deepfakes, audio deepfakes, and cheap fakes to test the effect of single modality against multimodality and how it affects individuals\' perceived claim accuracy and sharing intentions. In addition, the impact of cognitive ability on perceived claim accuracy and sharing intentions between conditions are also examined. The results suggest that individuals are likelier to perceive video deepfakes as more accurate than cheap fakes, but not audio deepfakes. Yet, individuals are more likely to share video deepfakes than cheap and audio deepfakes. We also found that individuals with high cognitive ability are less likely to perceive deepfakes as accurate or share them across formats. The findings emphasize that deepfakes are not monolithic, and associated modalities should be considered when studying user engagement with deepfakes.

Thousands of videos are posted on websites and social media every day, including Twitter, Facebook, WhatsApp, Instagram, and YouTube. Newspapers, law enforcement publications, criminal investigations, surveillance systems, Banking, the museum, the military, imaging in medicine, insurance claims, and consumer photography are just a few examples of places where important visual data may be obtained. Thus, the emergence of powerful processing tools that can be easily made available online poses a huge threat to the authenticity of videos. Therefore, it\'s vital to distinguish between true and fake data. Digital video forgery detection techniques are used to validate and check the realness of digital video content. Deep learning algorithms lately sparked a lot of interest in the field of digital forensics, such as Recurrent Neural Networks (RNN), Deep Convolutional Neural Networks (DCNN), and Adaptive Neural Networks (ANN). In this paper, we give a soft taxonomy as well as a thorough overview of recent research on multimedia falsification detection systems. First, the basic knowledge needed to comprehend video forgery is provided. Then, a summary of active and passive video manipulation detection approaches is provided. Anti-forensics, compression video methods, datasets required for video forensics, and challenges of video detection approaches are also addressed. Following that, we presented an overview of deepfake, and the datasets required for detection were also provided. Also, helpful software packages and forensics tools for video detection are covered. In addition, this paper provides an overview of video analysis tools that are used in video forensic applications. Finally, we highlight research difficulties as well as interesting research avenues. In short, this survey provides detailed information and a broader investigation to extract data and detect fraud video contents under one umbrella.

With the development of generative models, abused Deepfakes have aroused public concerns. As a defense mechanism, face forgery detection methods have been intensively studied. Remote photoplethysmography (rPPG) technology extract heartbeat signal from recorded videos by examining the subtle changes in skin color caused by cardiac activity. Since the face forgery process inevitably disrupts the periodic changes in facial color, rPPG signal proves to be a powerful biological indicator for Deepfake detection. Motivated by the key observation that rPPG signals produce unique rhythmic patterns in terms of different manipulation methods, we regard Deepfake detection also as a source detection task. The Multi-scale Spatial-Temporal PPG map is adopted to further exploit heartbeat signal from multiple facial regions. Moreover, to capture both spatial and temporal inconsistencies, we propose a two-stage network consisting of a Mask-Guided Local Attention module (MLA) to capture unique local patterns of PPG maps, and a Temporal Transformer to interact features of adjacent PPG maps in long distance. Abundant experiments on FaceForensics + + and Celeb-DF datasets prove the superiority of our method over all other rPPG-based approaches. Visualization also demonstrates the effectiveness of the proposed method.