AI深度伪造治理中的数字人格标识权益保护研究——以人脸、声纹的确权存证与检察公益诉讼为中心

… rights and evade the public trust while operating in a grey zone. This study advocates for recognition of a distinct ‘Right to Digital Persona’ … the malicious misuse of deepfake technology. …

… stars and digital creators show how personality rights are being tested against Al tools, deepfake filters, and viral content practices. This research examines whether personality rights in …

… deepfake technology necessitates an examination of the right of publicity and UK persona protection measures to determine what actions individuals can take to combat deepfakes. …

Unauthorised deepfakes are deeply problematic, from the spreading of misinformation to non‐consensual pornographic content. This paper asks whether deepfakes, digital replicas and human digital twins justify personality rights. To address this question, it examines the harms that deepfakes can cause through disinformation, demeaning content and displacing creative workers. It demonstrates that the current UK legal patchwork of passing off, intellectual property, defamation, and criminal laws do not adequately address these harms. Therefore, it proposes the introduction of personality rights into UK law, in the form of an automatic unwaivable personality right for 70 years after the death of the person, with appropriate exceptions to protect freedom of expression. Deepfakes are the hinges on which to open the door of personality rights in the UK, for protection against the harms of unauthorised digital replicas.

… stake, for our personal rights with the changes in technology. … Now, according to the law, we all have a right to privacy, … For this topic, however, we will be focusing on the right of publicity …

… of content producers and distributors end, and where the rights of individuals whose … that Deepfake content involving the likenesses of real individuals violates the rights of these …

… Personality and publicity rights deal with an individual's control over … rights due to deepfake usages and also seeks to recommend certain measures which can safeguard these rights …

… Evolving deepfake technology with its easy access and seeping of few strands of that technology in various social interaction applications has made it possible that personality …

This paper critically examines Denmark’s proposed amendment to its Copyright Act, particularly section 73(a), which aims to grant individuals intellectual property protection against the unauthorized sharing of realistic, digitally generated imitations of their physical traits (deepfakes). While recognizing the well-intentioned aim, this study contends that the Danish proposal is fundamentally flawed both conceptually and teleologically. The analysis demonstrates that copyright law is an inappropriate framework for safeguarding elements of personal identity. A significant teleological mismatch exists: copyright law promotes economic and cultural objectives by encouraging the creation of works, whereas personality rights are grounded in the principle of human dignity. This misalignment risks turning intrinsic personality traits into commodities and undermining the coherence of the copyright system. The study proposes that Spain’s Organic Act 1/1982 on the civil protection of the right to honour, privacy and one’s own image offers a more suitable alternative. Despite its origins in the 1980s, this act’s substantive and procedural design effectively addresses technological challenges such as deepfakes without requiring major reforms. The paper concludes that reinforcing existing civil protection mechanisms provides a more consistent solution than relying on copyright law.

<jats:p>.</jats:p>

… the future of deepfakes. Considering the centrality of personality rights to deepfakes, this chapter also explores the potential foundations of an EU-wide personality rights framework. It …

… rights’, we mean legal rights which protect a natural person’s interests in controlling their own identity, persona… seeking to invoke their right of publicity against deepfakes may be able to …

This research paper discusses the challenges of digital evidence usage in Deepfake crimes era in both the Egyptian and US Legislations. There is no doubt about the importance of keeping pace with the Law with behaviors that pose a threat to fundamental interests that deserve protection, especially in an era when information technology is instantaneously accelerating towards the creation of many modern technologies that raise many concerns, since artificial intelligence algorithms have helped to think about a large number of issues that did not exist a few years ago, such as the ease of processing big data and simultaneous machine translation, and one of those algorithms is Deepfake, which was classified as the most dangerous among artificial intelligence algorithms on Cybersecurity threats. With the complexity of Investigations of computer related crimes, due to the obstacles in gathering the evidence. The researcher seeks, after discussing the essence of digital evidence, stating its types, forms, characteristics, sources, principles, and challenges facing its application, as well as comparing between the laws regulating the digital evidence nationally, internationally concerning (Budapest convention) and The US federal rules of digital evidence, to present and set recommendations  to reduce the risks and challenges of these crimes, and to assist the legislator in addressing the shortcomings in Egyptian laws. Keywords: Digital evidence, Deepfake, Cybercrime, Digital privacy & Cybersecurity.

… evidence, even firsthand, is no longer considered a fact in the age of deepfakes. GAO SCI. TECH. … articulate reliable ways to authenticate digital evidence in the age of deepfakes, the …

In recent years, DeepFake is becoming a common threat to our society, due to the remarkable progress of generative adversarial networks (GAN) in image synthesis. Unfortunately, existing studies that propose various approaches, in fighting against DeepFake and determining if the facial image is real or fake, is still at an early stage. Obviously, the current DeepFake detection method struggles to catch the rapid progress of GANs, especially in the adversarial scenarios where attackers can evade the detection intentionally, such as adding perturbations to fool the DNN-based detectors. While passive detection simply tells whether the image is fake or real, DeepFake provenance, on the other hand, provides clues for tracking the sources in DeepFake forensics. Thus, the tracked fake images could be blocked immediately by administrators and avoid further spread in social networks. In this paper, we investigate the potentials of image tagging in serving the DeepFake provenance tracking. Specifically, we devise a deep learning-based approach, named FakeTagger, with a simple yet effective encoder and decoder design along with channel coding to embed message to the facial image, which is to recover the embedded message after various drastic GAN-based DeepFake transformation with high confidence. The embedded message could be employed to represent the identity of facial images, which further contributed to DeepFake detection and provenance. Experimental results demonstrate that our proposed approach could recover the embedded message with an average accuracy of more than 95% over the four common types of DeepFakes. Our research finding confirms effective privacy-preserving techniques for protecting personal photos from being DeepFaked.

THE CAPTURE is a mystery thriller series, that completed its second season on Peacock and BBC One. The British television drama revolves around the alteration of direct audiovisual evidence on the command of a special unit that believes there is enough circumstantial evidence to either convict or acquit an individual of a felony. Based on the plot of the television series, this paper explores the potential for a variety of AI-enabled applications to be used in the course of criminal proceedings. The implications of evidence tampering are considered through AI manipulation toward the realization that deepfake evidence may well be admitted in court dependent on the human decision-maker. Will the future demand the interpolation of visual evidence for high profile criminal cases, and what does the existence of Generative AI and deepfakes mean for the forensic analysis of audiovisual evidence? After contemplating the socio-technical plausibility of the central premise of THE CAPTURE, this paper then turns to its legal implications. Drawing on examples from U.S. and Australian legal frameworks, the paper considers the consequences of AI-corrected, augmented or generated audiovisual evidence on three facets of natural justice: (1) the presumption of innocence; (2) the fair trial; and (3) lawyers’ ethical duties of competence and to the administration of justice. The key takeaways of the paper are that: (1) deepfake evidence will continue to proliferate; (2) that the law will need to address both the substantive and procedural impacts of such evidence, and (3) that the legal profession must continue to educate its lawyers and practitioners, and associated stakeholders, of the nature, uses and risks posed by deepfake audiovisual artefacts to maintain public trust in the legal system.

The current problem in the world of deepfakes faces the challenge of a 700% increase in fraud cases in the fintech sector (2023) with projected losses reaching $40 billion by 2027. Although intensive research focuses on model architecture, the preprocessing stage, which is fundamental and crucial in the preservation of manipulation artifacts, is still under-explored. This study presents a systematic comparative analysis of preprocessing techniques for deepfake detection using EfficientNetV2-S as the backbone on the Celeb-DF v2 dataset (5,639 videos, 1:1 real-to-fake ratio). We evaluate two scenarios: (1) baseline preprocessing with BlazeFace face detection, photometric normalization, and light augmentation; (2) adaptive multi-scale enhancement that integrates gamma correction, JPEG compression simulation, adaptive sharpening-blur (CPU-side), with Symlet-4 level-3 wavelet decomposition and spatial attention fusion (GPU-side). The experimental results show significant improvements: at the frame level, AUROC increased from 99.07% to 99.62% with a 59.2% reduction in false positives (240→98 cases); at the video level, the proposed method achieved perfect precision of 100% on the Celeb-DF v2 test set (zero false positives out of 83 real videos). This result is benchmark-specific and does not imply universal perfect precision, AUROC of 99.96%, and Equal Error Rate of 0.60%. Error analysis identified that detection failures mainly occurred under extreme lighting conditions and high-quality deepfakes with adversarial post-processing. This research advances the field through: (1) a systematic evaluation framework that rigorously quantifies the impact of preprocessing, (2) an adaptive multi-scale pipeline that preserves and amplifies forensic artifacts without introducing additional backbone parameters, and (3) empirical evidence that preprocessing significantly elevates deepfake detection performance. The study is limited by its evaluation on a single dataset and its use of empirically determined fusion weights, which motivate future work on cross-dataset validation, temporal modeling, and learnable fusion mechanisms.

Deepfake technology poses a profound challenge to the integrity of facial evidence in criminal justice, threatening the authenticity and admissibility of such evidence in the courtroom. In this research, a specialized deepfake detection system tailored for facial evidence verification was developed, aiming to counteract the influence of deepfake technology. The proposed system integrates a unique combination of video-frame selection, confidence thresholds, prediction timestamps, and heat maps for individual frames of suspect videos. This methodological fusion is designed to support forensic analysts by enhancing the reliability and trustworthiness of video evidence used in judicial settings. Our comprehensive evaluation involved diverse user groups participating in experimental scenarios to assess the effectiveness of the system. The results indicated that the combined features of the system significantly enhanced the detection of fabricated evidence, fostering high levels of confidence and trust among users. Moreover, this study delves into the legal and ethical considerations surrounding the deployment of deep fake-detection technologies, underscoring the necessity for legal frameworks to evolve in response to emerging digital threats. By addressing both the technical and jurisprudential challenges, this research contributes to safeguarding the evidential value of facial recognition in the judicial process against the disruptive potential of deepfake technologies.

… It will start by collecting and preserving evidence with the main emphasis put on obtaining raw files and calculating cryptographic hashes to confirm integrity of documents [12]. In the …

… This fundamental change shows that the crisis caused by deepfakes is not only forensic, but … To understand the depth of the challenge posed by documentary deepfakes, three …

… Legal literature and academic scholarship are just beginning to examine the social and legal ramifications of deepfakes,5 and deepfake evidence in court proceedings is a new and …

This research focuses on the detection of deepfake images and videos for forensic analysis using deep learning techniques. It highlights the importance of preserving privacy and authenticity in digital media. The background of the study emphasizes the growing threat of deepfakes, which pose significant challenges in various domains, including social media, politics, and entertainment. Current methodologies primarily rely on visual features that are specific to the dataset and fail to generalize well across varying manipulation techniques. However, these techniques focus on either spatial or temporal features individually and lack robustness in handling complex deepfake artifacts that involve fused facial regions such as eyes, nose, and mouth. Key approaches include the use of CNNs, RNNs, and hybrid models like CNN-LSTM, CNN-GRU, and temporal convolutional networks (TCNs) to capture both spatial and temporal features during the detection of deepfake videos and images. The research incorporates data augmentation with GANs to enhance model performance and proposes an innovative fusion of artifact inspection and facial landmark detection for improved accuracy. The experimental results show near-perfect detection accuracy across diverse datasets, demonstrating the effectiveness of these models. However, challenges remain, such as the difficulty of detecting deepfakes in compressed video formats, the need for handling noise and addressing dataset imbalances. The research presents an enhanced hybrid model that improves detection accuracy while maintaining performance across various datasets. Future work includes improving model generalization to detect emerging deepfake techniques better. The experimental results reveal a near-perfect accuracy of over 99% across different architectures, highlighting their effectiveness in forensic investigations.

… in cases involving suspected deepfakes. The article … of expert examinations involving deepfake evidence. It argues that courts … Preservation: What measures were taken to preserve the …

Technology has deeply impacted society and how activities are conducted. This is not different from the development of deepfake technology. Deepfake is of serious concern to the public generally, but more particularly, it is of grave concern to the admission of evidence at trial. This paper espouses deepfake technology's inherent challenges to the Federal Rules of Evidence in the United States. As the use of this technology and its dangerous tendencies expands, it continues to bear on its multiplier effects in the judicial system concerning the authenticity of pieces of evidence in the courtroom. The paper argues that in the wake of this unprecedented development of deepfake technology, capable of manipulating evidence and misleading the courts, courts must deploy extra measures to ensure that evidence presented before it is the original. Thus, the paper proposes measures that courts may implement to address the challenges of deepfake.

This paper focuses on the criminal law response to AI fraud crimes and systematically analyzes the theoretical and practical dilemmas in terms of the responsible party, behavior determination, subjective intent, causality, etc. The study points out that traditional criminal law has application difficulties in the face of problems such as multi-subject participation, behavioral independence and ambiguous intent brought about by AI technology. The article presents a "compromise" view, advocating the combination of technological governance and criminal law regulation to construct a four-in-one criminal law regulation path: improving criminal legislation, clarifying judicial responsibilities and evidence rules, promoting social co- governance, and using AI technology to combat fraud. The conclusion suggests that effective regulation of AI fraud should be achieved through the combined efforts of law, technology and social governance.

Results: Mainland China faces three key issues: misprioritizing public morality over sexual autonomy, insurmountable private prosecution evidence burdens (relief rate <30%), and inconsistent convictions. Taiwan Region’s specialized charge achieves 70%+ relief rates via public prosecution and targeted protection. Conclusions: Mainland China must establish a specialized DFII charge—centered on sexual autonomy, using public prosecution—to meet UN obligations, resolve judicial dilemmas, and advance cybercrime governance. Future research may explore unidentifiable DFII.

… of IBSA (including Deepfakes) in England and under US Federal Criminal Law in Section 3. … in a public or commercial setting. Thirdly the depiction must not be a matter of public concern …

With the rapid advancement of information technology, electronic data has become increasingly integrated into all aspects of social life and judicial practice. As a new type of evidence, electronic data plays a crucial role in criminal, civil, and administrative proceedings. Yet, its application raises profound challenges concerning legality, procedural efficiency, and standardization. This paper explicitly situates itself within procedural law, evidentiary reform, and judicial governance, highlighting its dual contribution to theory and practice. Methodologically, it adopts literature review, comparative analysis, case studies, doctrinal research, and selected empirical data. The study identifies three core challenges: insufficient legal legitimacy of examination procedures, procedural inefficiency caused by technological and institutional fragmentation, and the absence of nationwide technical standards. Drawing on international experiences—particularly from the United States, the European Union, ASEAN, and selected Asian jurisdictions—the paper proposes a framework to enhance legality, efficiency, and cross-border cooperation. The findings enrich the discourse on procedural justice in the digital age and provide targeted recommendations for strengthening China’s evidentiary system and smart court development.

This article considers the prominent private law and regulatory problems relating to the practice of AI ‘resurrection’ where a digital replica with the appearance, voice, and personality of the deceased is created to interact with the real world. The intention is to offer solace and comfort to the bereaved by interacting with a digital representation of their loved one, providing a semblance of continued presence as if the departed had never left this world. This article first considers the industry of AI ‘resurrection’ in the context of ‘grief tech’. It uses the Chinese experience for discussion as to date the practice is most prevalent in China but the legal considerations are applicable to other jurisdictions. It then discusses whether and how the practice may constitute the infringement of the deceased’s portrait right as a personality interest protected under the Civil Code and its remedies. It finally considers the regulatory aspects of AI ‘resurrection’ with respect to personal information, deep synthesis and AI generative technology.

… liable for restricting access to deepfakes. Moreover, these … private and public interests as well as private and public regulatory … courts and People's Procuratorates, and administrative …

… handled over 6,700 public interest litigation cases related to … have been ramping up deepfake offerings to carry out virtual … and the procurator-generals of procuratorates are selected …

… ignores the independence of legal interest in infringing on deepfake behaviour and the … law or public order and good customs. Public order and good customs are types of public interest…

… video or a replayed recording or a deepfake," said Andrew Bud, … handled over 6,700 public interest litigation cases related to … and the procurator-generals of procuratorates are selected …

In recent years, industry leaders and researchers have proposed to use technical provenance standards to address visual misinformation spread through digitally altered media. By adding immutable and secure provenance information such as authorship and edit date to media metadata, social media users could potentially better assess the validity of the media they encounter. However, it is unclear how end users would respond to provenance information, or how to best design provenance indicators to be understandable to laypeople. We conducted an online experiment with 595 participants from the US and UK to investigate how provenance information altered users' accuracy perceptions and trust in visual content shared on social media. We found that provenance information often lowered trust and caused users to doubt deceptive media, particularly when it revealed that the media was composited. We additionally tested conditions where the provenance information itself was shown to be incomplete or invalid, and found that these states have a significant impact on participants' accuracy perceptions and trust in media, leading them, in some cases, to disbelieve honest media. Our findings show that provenance, although enlightening, is still not a concept well-understood by users, who confuse media credibility with the orthogonal (albeit related) concept of provenance credibility. We discuss how design choices may contribute to provenance (mis)understanding, and conclude with implications for usable provenance systems, including clearer interfaces and user education.

… , access control, provenance, transparency, and evidence verification presents several … digital evidence throughout the forensic investigation. We only insert or delete digital evidence …

… , it is difficult to audit the log for traceability and provenance if a user decides to be malicious. To … scenario of evidence actions within TheHive security incident response platform (SIRP). …

Generative AI (GenAI) now produces text, images, audio, and video that can be perceptually convincing at scale and at negligible marginal cost. While public debate often frames the associated harms as “deepfakes” or incremental extensions of misinformation and fraud, this view misses a broader socio-technical shift: GenAI enables synthetic realities—coherent, interactive, and potentially personalized information environments in which content, identity, and social interaction are jointly manufactured and mutually reinforcing. We argue that the most consequential risk is not merely the production of isolated synthetic artifacts, but the progressive erosion of shared epistemic ground and institutional verification practices as synthetic content, synthetic identity, and synthetic interaction become easy to generate and hard to audit. This paper (i) formalizes synthetic reality as a layered stack (content, identity, interaction, institutions), (ii) expands a taxonomy of GenAI harms spanning personal, economic, informational, and socio-technical risks, (iii) articulates the qualitative shifts introduced by GenAI (cost collapse, throughput, customization, micro-segmentation, provenance gaps, and trust erosion), and (iv) synthesizes recent risk realizations (2023–2025) into a compact case bank illustrating how these mechanisms manifest in fraud, elections, harassment, documentation, and supply-chain compromise. We then propose a mitigation stack that treats provenance infrastructure, platform governance, institutional workflow redesign, and public resilience as complementary rather than substitutable, and outline a research agenda focused on measuring epistemic security. We conclude with the Generative AI Paradox: as synthetic media becomes ubiquitous, societies may rationally discount digital evidence altogether, raising the cost of truth for everyday life and for democratic and economic institutions.

… The paper also engages with scholarship on digital evidence provenance, interview … Additionally, the report failed to conduct digital persona verification— determining whether social …

Abstract In the face of escalating cyber threats, a real-time automated security evidence collection system for cloud-based digital forensics investigations is essential for identifying and mitigating malicious activities. However, the substantial volumes of data generated by modern cloud-based digital systems pose difficulties in collecting and analyzing evidence promptly and systematically. To address these challenges, this research introduces an architecture that combines a security lake and a modern data lake. The primary objective of this architecture is to overcome the obstacles associated with gathering evidence from multiple cloud-based accounts and regions while ensuring the flexibility and scalability required to manage the ever-expanding data volumes encountered in cloud-based digital forensics investigations. This work focuses on gathering security events from multiple accounts and regions within a cloud environment in real-time while maintaining the integrity of the evidence and storing them in lakes, providing investigators with the flexibility to move between these lakes for analysis to get quick results. This is achieved through the utilization of security lake and modern data architecture. To validate the system, we tested it within a university system comprising numerous accounts spread across different regions within an AWS environment. Overall, the proposed system effectively gathers evidence from various sources and consolidates all data lakes into a single account. These lakes were then utilized for analyzing the evidence using Athena and Wazuh.

In this paper we present the details of the multilevel liveness verification (MLLV) framework proposed for realizing a secure face-voice biometric authentication system that can thwart different types of audio and video replay attacks. The proposed MLLV framework based on novel feature extraction and multimodal fusion approaches, uncovers the static and dynamic relationship between voice and face information from speaking faces, and allows multiple levels of security. Experiments with three different speaking corpora VidTIMIT, UCBN and AVOZES shows a significant improvement in system performance in terms of DET curves and equal error rates (EER) for different types of replay and synthesis attacks.

Objective: to specify the models of legal regulation in the sphere of biometric identification and authentication with facial recognition technology in order to elaborate recommendations for increasing information security of persons and state-legal protection of their right to privacy.Methods: risk-oriented approach in law and specific legal methods of cognition, such as comparative-legal analysis and juridical forecasting, are significant for the studied topic and allow comparing the legal regulation models used in foreign countries and their unions in the sphere of biometric identification and authentication with facial recognition systems, forecasting the possible risks for the security of biometric data, taking into account the prospects of further dissemination of the modern facial recognition technology, and to shape recommendations on legal protection of biometric data.Results: the ways are proposed to further improve legislation of the Republic of Kazakhstan and other countries currently developing the legal regulation of biometric data, regarding the admissible criteria for using the facial recognition technology, the elaboration of categorization of biometric systems with a high and low risk levels (by the example of the experience of artificial intelligence regulation in the European Union), and the necessity to introduce a system of prohibitions of mass and unselective surveillance of humans with video surveillance systems, etc.Scientific novelty: consists in identifying a positive advanced foreign experience of developing legal regulation in the sphere of facial recognition based on biometry (European Union, the United States of America, the United Kingdom of Great Britain and Northern Ireland), which can be used for further improvement of the national legislation in order to create more effective mechanisms of legal protection of personal data, including biometric information.Practical significance: based on risk-oriented approach and comparative analysis, the research allows elaborating measures for enhancing the legal protection of biometric data and ensuring effective protection of civil rights and freedoms by forecasting further expansion of the modern facial recognition technology.

Biometric authentication is a fast-growing, novel technology that makes the identity verification process secure and user-friendly with unique physiological and behavioral indicators. The paper attempts to meet a theory on biometric systems' development, hazards and constraints by examining various biometric technologies like fingerprint, facial acknowledgement, iris checking, and behavioral biometrics. The integration of machine learning and artificial intelligence has made these systems more accurate and reliable, and hence, they can be used on mobile devices, banks, and border control. Since biometric authentication systems are beneficial, there are many risks, including privacy, data breaches, spoofing attacks and regulatory issues. While these risks are unavoidable as more and more organizations are embracing biometric systems, it is only right that data protection is strong, that legal frameworks are abided by, and that the practice is ethical. The paper explains the need to balance security and user convenience that takes care of the secure storage and transmission of biometric data and maintains compliance with evolving regulations. In addition, it stresses the importance of repeatedly observing and updating the biometric authentication systems to guarantee their integrity and security. Biometric authentication is poised to become a powerful tool for increasing security in many sectors. However, this success depends largely on considering different technological, legal and ethical factors to diminish any risks such a system might present.

Identity theft is one of the fastest-growing forms of cybercrime, driven by large-scale data breaches, phishing, and increasingly sophisticated impersonation attacks. Traditional identity verification methods such as passwords, PINs, and physical documents have proven inadequate in ensuring security at scale. Artificial Intelligence (AI) has emerged as a transformative enabler of next-generation identity verification by leveraging multimodal techniques, including facial recognition, voice biometrics, and document authentication. The paper discusses how AI- based verification systems can be used to prevent identity theft and how the system is used in real-time adaptive, and frictionless authentication over high-stakes areas, including banking, healthcare, e-commerce, and government services. We introduce a multi-layered verification system that combines the facial, voice and document verification modules in a single decision layer to minimize the false positives and negative but enhances the system resistance to spoofing and adversarial attacks. Practical implementations, advantages and governance are described using case studies of financial institutions, e-commerce websites and national identity programs. Nevertheless, there are still obstacles, such as demographic bias, privacy risks, adversarial vulnerability and lack of a coherent regulatory framework that makes it difficult to achieve mass adoption. In the future, we will address future directions in the area of decentralized identity, federated learning, zero-knowledge proofs, explainable AI, and international regulatory alignment. These innovations will work towards building trust, fairness and interoperability in digital identity ecosystems. Finally, this paper shows that AI-based identity verification is not merely a technological breakthrough but one of the essential needs to protect individuals, organizations, and governments against identity theft during the digital age.

… synthetic media proposals and laws mandating watermarking of generative AI (GenAI) content (eg, “AI … to the speaker and to society that outweighs public interest in disclosure.”Justice …

In recent years, artificial intelligence has been deployed by online platforms to prevent the upload of allegedly illegal content or to remove unwarranted expressions. These systems are trained to spot objectionable content and to remove it, block it, or filter it out before it is even uploaded. Artificial intelligence filters offer a robust approach to content moderation which is shaping the public sphere. This dramatic shift in norm setting and law enforcement is potentially game-changing for democracy. Artificial intelligence filters carry censorial power, which could bypass traditional checks and balances secured by law. Their opaque and dynamic nature creates barriers to oversight, and conceals critical value choices and tradeoffs. Currently, we lack adequate tools to hold them accountable. This paper seeks to address this gap by introducing an adversarial procedure— – Contesting Algorithms. It proposes to deliberately introduce friction into the dominant removal systems governed by artificial intelligence. Algorithmic content moderation often seeks to optimize a single goal, such as removing copyright-infringing materials or blocking hate speech, while other values in the public interest, such as fair use or free speech, are often neglected. Contesting algorithms introduce an adversarial design which reflects conflicting values, and thereby may offer a check on dominant removal systems. Facilitating an adversarial intervention may promote democratic principles by keeping society in the loop. An adversarial public artificial intelligence system could enhance dynamic transparency, facilitate an alternative public articulation of social values using machine learning systems, and restore societal power to deliberate and determine social tradeoffs.

… With the widespread availability of Artificial Intelligence (AI) tools, specifically Generative AI, … by Generative AI. We must confront two possibilities: first, that evidence presented is AI …

… does it automatically override core public interests, particularly freedom of expression31. … In the digital and artificial intelligence context, where reproduction and imitation increasingly …

Deepfake technology, fuelled by advances in artificial intelligence (AI), poses an escalating threat to biometric authentication systems. This study evaluates whether AI-generated deepfakes can bypass facial recognition security in Apple Face ID, Android Face Unlock and Windows Hello, with additional analysis of their implications for Microsoft Authenticator and Azure Active Directory. Using three open-source tools: Deep-FaceLab, FaceSwap and FaceFusion controlled experiments were conducted under standardised playback conditions to simulate realistic spoofing attempts. Detection performance was assessed with commercial AI-based forensic tools. The results reveal wide variation in system resilience: some configurations resisted all attacks, while others were compromised when playback was optimised. Critically, a full attack chain was demonstrated in which Apple Face ID (iPhone 11, iOS 18.6) was spoofed, enabling bypass of multi-factor authentication (MFA) and escalation to Azure resources. Detection tools showed mixed effectiveness, with consistent failures against certain outputs, underscoring the need for integrated and adaptive defences. This paper makes three contributions: (1) it provides an empirical assessment of deepfake spoofing across major biometric systems under realistic conditions; (2) it demonstrates a complete MFA bypass via deepfake-driven spoofing; and (3) it delivers a comparative evaluation of forensic detection tools, highlighting both their strengths and limitations. Together, these findings emphasise the importance of advancing liveness detection, adaptive anti-spoofing measures, and forensic integration to sustain trust in biometric authentication.

Recent research has demonstrated that lip-based speaker authentication systems can not only achieve good authentication performance but also guarantee liveness. However, with modern DeepFake technology, attackers can produce the talking video of a user without leaving any visually noticeable fake traces. This can seriously compromise traditional face-based or lip-based authentication systems. To defend against sophisticated DeepFake attacks, a new visual speaker authentication scheme based on the deep convolutional neural network (DCNN) is proposed in this paper. The proposed network is composed of two functional parts, namely, the Fundamental Feature Extraction network (FFE-Net) and the Representative lip feature extraction and Classification network (RC-Net). The FFE-Net provides the fundamental information for speaker authentication. As the static lip shape and lip appearance is vulnerable to DeepFake attacks, the dynamic lip movement is emphasized in the FFE-Net. The RC-Net extracts high-level lip features that discriminate against human imposters while capturing the client’s talking style. A multi-task learning scheme is designed, and the proposed network is trained end-to-end. Experiments on the GRID and MOBIO datasets have demonstrated that the proposed approach is able to achieve an accurate authentication result against human imposters and is much more robust against DeepFake attacks compared to three state-of-the-art visual speaker authentication algorithms. It is also worth noting that the proposed approach does not require any prior knowledge of the DeepFake spoofing method and thus can be applied to defend against different kinds of DeepFake attacks.

… This study provides a thorough evaluation of four distinguished deepfake detection tools—… deployable deepfake detection systems in the context of real-world media authentication. …

The recent advent of extremely realistic deepfakes generated by GANs and diffusion models has presented an unsparing challenge to face-based authentication and remote identity verification. Tradition Single-mode detectors are lost during compression, after-processing, and to unseen generator families and thus hurt inauthentic that systems via imitation. In this paper, we propose Multimodal Vision-Transformer Forensics (MViT-Forensics), a transformer-based forensic framework luxurious for secure authentication applied in deepfake Scenario. MViT-Forensics processes spatial RGB embeddings (such as Swin/ViT backbone), frequency domain residuals (DCT/FFT process step) and physiological rPPG waveforms (attention-based temporal encoder) together, and combines them using a cross-attention multimodal transformer. The above anomaly-aware intrusion scoring head will generate a confidence score in the form of authenticity level for an authentication policy. We analyze the approach on four public benchmarks—FaceForensics++, Celeb-DF v2, DFDC Preview and the Multimodal DeepFake Dataset (MMDF)— in standard and compressed conditions. Our model reaches average accuracies of 98.1% (FF++), 96.4% (Celeb-DF), 94.7% (DFDC) and 97.2% (MMDF) with mean AUC $\approx 0: 99$, and maintains $>90$ percent accuracy at high levels of compression while having better performance than strong CNN or transformer baselines. Ablations verify the complementary contributions in frequency and physiology branches and the superiority of cross-attention fusion. Our findings suggest that multimodal attention-based forensics with intrusion scoring could close the gap between academic detectors and operational authentication needs.

With the rapid advancement of visual generative models such as Generative Adversarial Networks (GANs) and stable Diffusion, the creation of highly realistic Deepfake through automated forgery has significantly progressed. This paper examines the advancements in Deepfake detection and defense technologies, emphasizing the shift from passive detection methods to proactive digital watermarking techniques. Passive detection methods, which involve extracting features from images or videos to identify forgeries, encounter challenges such as poor performance against unknown manipulation techniques and susceptibility to counter-forensic tactics. In contrast, proactive digital watermarking techniques embed specific markers into images or videos, facilitating real-time detection and traceability, thereby providing a preemptive defense against Deepfake content. We offer a comprehensive analysis of digital watermarking-based forensic techniques, discussing their advantages over passive methods and highlighting four key benefits: real-time detection, embedded defense, resistance to tampering, and provision of legal evidence. Additionally, the paper identifies gaps in the literature concerning proactive forensic techniques and suggests future research directions, including cross-domain watermarking and adaptive watermarking strategies. By systematically classifying and comparing existing techniques, this review aims to contribute valuable insights for the development of more effective proactive defense strategies in Deepfake forensics.

… These manipulations can range from easy edits to sophisticated deepfake … forensics and deep learning. A visual representation of keywords in deep learning for video authentication …

The progress of artificial intelligence (AI) has enabled the creation of very realistic synthetic media, also known as deepfakes, which poses a serious threat to information integrity and social confidence. The article examined the process of detecting and authenticating deep fakes using hybrid AI models. The researchers employed the case study methodology, based on the Celeb-DF V2 dataset, one of the most challenging datasets for generating high-quality manipulated videos. The suggested system combined convolutional neural networks (CNNs) to extract spatial features, recurrent neural networks (LSTMs/GRUs) to model temporal consistency, and transformer systems to analyse fine-grained context. The researchers bundled these parts together to enhance robustness and generalisation in an ensemble mechanism. They also introduced provenance tracking and semi-fragile watermarking to supplement detection, enabling proactive authentication and watermark verification of media through blockchain-based provenance tracking. The experimental findings showed that the hybrid models were more accurate, achieved higher F1 Scores, and were more robust to adversarial manipulations than the single-model baselines. The hybrid with a transformer achieved the best accuracy (0.95 AUC) and the lowest false-positive rate (6%), but at the expense of slower processing speeds. Authentication tools also helped strengthen trust by verifying the originality of content and flagging potential manipulation before it was classified. The results have revealed that hybrid AI models, when implemented with authentication strategies, represent a more effective and legitimate approach to addressing the threats of misinformation, fraud, and loss of trust among the population in the face of deepfakes.

Deepfake-enabled identity abuse has moved from a peripheral cyber-risk to an operational threat for U.S. banking, especially in remote onboarding, account recovery, contact-center authentication, and high-risk payment authorization. The problem is no longer limited to obvious synthetic media. Financial institutions increasingly face blended attacks that combine forged identity documents, face or voice cloning, injected video streams, social-engineering pressure, mule accounts, and adaptive retries across channels. This paper develops a publication-ready research framework for machine learning systems that identify deepfake-driven customer impersonation, authentication evasion, and identity abuse in U.S. banking environments. The study synthesizes regulatory guidance, public fraud data, biometric spoofing research, deepfake detection literature, and banking fraud analytics to propose a multimodal detection architecture spanning document forensics, face presentation attack detection, deepfake-video detection, speaker anti-spoofing, behavioral biometrics, device and network telemetry, graph-based entity resolution, and risk-calibrated decisioning. Real public evidence motivates the design: FTC data show reported fraud losses reached $12.5 billion in 2024, including $2.95 billion in imposter-scam losses, while FinCEN reported increasing suspicious activity narratives involving deepfake media targeting financial institutions. The proposed methodology treats identity abuse as a sequential, multimodal, and adversarial classification problem rather than a single-screen biometric check. The paper argues that the most effective defense is not one detector but an explainable ensemble with smart friction, human escalation, and governance aligned to NIST identity-proofing standards, anti-money-laundering expectations, and consumer-protection obligations. By linking technical detection with operational controls, the study provides a practical blueprint for authentication, resilient fraud prevention, and more trustworthy digital banking systems.