Abstract
Artificial intelligence has transformed identity from a primarily embodied and relational phenomenon into a distributed object of continuous collection, classification, prediction, and synthetic reproduction. This transformation does not support the simplistic conclusion that persons possess an absolute property right over every representation, judgment, or fact associated with them. It supports a more defensible bundle of qualified moral rights grounded in autonomy, dignity, privacy, fairness, non-domination, bodily integrity, and the social conditions required for self-development. This paper argues that persons possess moral rights against unauthorized synthetic replacement, indiscriminate biometric observation, harmful inferential appropriation, false attribution, substituted agency, and institutional practices that close the future by fastening individuals to permanent, searchable profiles. It further argues that digital consent frequently fails as a justificatory mechanism because formal agreement occurs under conditions of opacity, asymmetrical power, practical necessity, and unknowable downstream use. Deepfakes, automated surveillance networks, predictive profiling, and adversarial privacy technologies reveal distinct stages of identity governance rather than variations of a single privacy problem. The appropriate normative response is therefore neither unrestricted technological use nor total individual control, but a structured identity ethics requiring necessity, proportionality, contextual legitimacy, contestability, and meaningful opportunities for resistance and correction.
Keywords: identity ethics, artificial intelligence, digital integrity, informational self-determination, deepfakes, biometric surveillance, inferential privacy, adversarial clothing
Introduction
Artificial intelligence has altered the conditions under which persons become socially intelligible. Identity is still experienced through consciousness, embodiment, memory, social relationships, and narrative continuity, but it is now also assembled through biometric profiles, behavioral records, digital avatars, location histories, and algorithmic classifications. These digital manifestations are not merely passive descriptions of a preexisting self. They increasingly influence how institutions recognize a person, what opportunities are presented or withheld, which risks are attributed to that person, and whether the person’s own speech, appearance, or conduct can be convincingly simulated. The resulting moral problem is therefore broader than privacy in the conventional sense. Privacy concerns boundaries of access, use, and disclosure; identity concerns the substantive standing of the person who is observed, represented, classified, and acted upon (Matar & Murray, 2025; Munch & Mainz, 2023).
The central question is not whether persons own identity as though it were a conventional commodity. That framing is attractive because property language supplies familiar rules of control, exclusion, transfer, and remedy, but it is conceptually incomplete. Identity is partly embodied, partly relational, partly socially constituted, and partly authored through time. A person may have legitimate claims against unauthorized appropriation without possessing a general power to control every public reference, critical judgment, historical record, or artistic representation involving that person. The more defensible position is that persons possess a bundle of moral rights over specific identity-making processes, particularly creation, observation, identification, prediction, manipulation, and downstream use (Shoemaker, 2008; Herring & Wall, 2017; Hamuľák et al., 2025).
This distinction matters because identity harms are cumulative but not interchangeable. Observation may violate contextual privacy without fabricating anything. Identification may destroy practical obscurity by linking otherwise scattered traces. Prediction may appropriate intimate information that was never directly disclosed. Synthetic creation may replace a person’s agency with fabricated speech or conduct. Manipulation may alter the conditions through which others perceive and respond to the person. A coherent identity ethics must therefore assess each stage independently before asking whether the combined system is morally justified (Mainz & Uhrenfeldt, 2021; Munch, 2021; Rumbold & Wilson, 2019).
The Distributed Self and the Body as Data
Personal identity is commonly analyzed through at least three overlapping domains. Personal identity concerns continuity and persistence across time; social identity concerns group membership, recognition, and relational position; narrative identity concerns the ongoing interpretation through which a person understands a life as a story still capable of revision. Artificial intelligence destabilizes all three domains because it converts bodily and behavioral attributes into portable data that can be copied, combined, classified, and operationalized outside the person’s immediate control (Olson, 2023; Shoemaker, 2008; Matar & Murray, 2025).
The concept of the body as data captures this transformation. Biometric sensors, facial-recognition systems, wearables, neural interfaces, and ordinary digital services generate information that is derived from the person but detached from the physical body. Once detached, the information can function as a digital prosthesis, an extension of personality that remains socially consequential even when the individual is absent. Facial geometry, voice, gait, typing patterns, health information, and location histories can become mechanisms of recognition, exclusion, persuasion, or simulation. Treating such material as ordinary content overlooks its embodied origin and its capacity to mediate the person’s practical agency (Bublitz, 2022; Hamuľák et al., 2025; Herring & Wall, 2017).
This does not mean that every datum generated by a body is identical to the body itself. Such an equivalence would erase necessary distinctions between physical injury, informational harm, reputational harm, and commercial appropriation. The relevant claim is narrower: some body-derived information is sufficiently constitutive of agency, recognizability, and social standing that its unauthorized extraction or manipulation can violate interests analogous to those protected by bodily integrity. Digital integrity is therefore best understood as an extension of personhood protection, not as a metaphorical assertion that data and flesh are literally the same thing (Patella-Rey, 2018; Hamuľák et al., 2025).
Why Consent Is an Inadequate Moral Foundation
Many legal and commercial systems attempt to justify data collection through consent.
The justificatory structure appears straightforward: an autonomous person receives information, understands the proposed use, and voluntarily authorizes it. In actual digital environments, however, each element of that model is routinely weakened. Terms are lengthy and technically obscure, access is conditioned on acceptance, systems evolve after the initial agreement, and downstream inferences are often impossible for the user to anticipate. Formal acceptance may therefore establish that a button was pressed, but it does not necessarily establish informed and voluntary authorization (Mainz & Uhrenfeldt, 2021; Véliz & Law, 2023).
The problem is not merely that users fail to read. That explanation individualizes a structural failure. Digital participation is frequently necessary for employment, education, communication, commerce, and access to public or quasi-public services. Under these conditions, a take-it-or-leave-it choice may be legally recognizable while remaining morally deficient. The power asymmetry is intensified when the data controller knows far more about the architecture, analytical capacity, and future commercial use of the system than the person who nominally consents. Consent then functions less as an exercise of autonomy than as a liability-shifting mechanism.
Artificial intelligence further destabilizes consent because the ethically significant act may occur after collection. A person may knowingly disclose a photograph without authorizing facial recognition, a voice recording without authorizing synthetic cloning, or a location trace without authorizing predictive risk classification. Public availability and prior disclosure do not entail a general waiver of claims against aggregation, inference, repurposing, or synthetic reproduction. The crucial moral question is not simply whether information was accessible, but whether the downstream operation respects the context, expectations, agency, and vulnerability of the person concerned (Rumbold & Wilson, 2019; Munch, 2021; Mainz, 2024).
A Bundle of Moral Rights over Identity
The literature supports six principal moral rights, each qualified by competing claims and institutional context. These rights should not be treated as isolated entitlements. Together, they describe the minimum conditions under which persons can remain authors and participants in the construction of their identities rather than raw material for external systems.
The Right Against Unauthorized Synthetic Identity Creation
Generative artificial intelligence can reproduce facial appearance, vocal style, gesture, and behavioral patterns with sufficient realism to create digital duplicates. The distinctive wrong is not exhausted by deception. Even an obvious deepfake can appropriate a person’s self-presentation by assigning speech or conduct to the person without authorization. The synthetic artifact substitutes fabricated agency for real agency, using the person as an instrument for the creator’s purposes and degrading the person’s authority over what they actually said or did (de Ruiter, 2021; Lundgren, 2026).
A useful distinction is the difference between reference and replacement. Reference places a person within commentary, criticism, parody, journalism, art, or historical narration. Replacement simulates the person’s presence, performance, or communication in a manner that bypasses the person’s agency. Because expression and public discourse require substantial freedom to refer to persons, identity ethics cannot establish an unlimited veto over all representations. It can, however, impose a strong presumption against synthetic replacement, particularly when the simulation is intimate, exploitative, deceptive, commercially appropriative, or likely to be treated as authentic (Chopra et al., 2025; Bosher, 2026).
The Right to Contextual Privacy and Practical Obscurity
Privacy is not equivalent to secrecy. A person may appear in public while retaining a legitimate expectation that the appearance will not be permanently identified, aggregated, and rendered searchable across time. Practical obscurity describes the friction that historically limited the scale and persistence of observation. Automated biometric systems remove that friction by converting fleeting presence into durable identification. The resulting moral change is qualitative, not merely quantitative, because continuous identifiability alters how persons move, associate, experiment, and dissent (Marmor, 2015; Macnish, 2018; Smith & Miller, 2021). Automated license plate readers and networked camera systems illustrate the problem.
A single observation of a vehicle in public may be morally unremarkable. A network that records repeated movements, links them to an individual, stores them, and enables retrospective or predictive analysis creates a different object: a behavioral profile. When such systems are deployed without meaningful public knowledge, municipal authorization, clearly defined retention limits, or accountable access controls, the practice becomes a form of identity capture rather than ordinary observation. Public-safety goals may justify narrowly tailored monitoring, but generalized surveillance requires evidence of necessity, proportionality, and institutional accountability, not rhetorical invocation of security (Smith & Miller, 2021; Wang et al., 2024).
The Right Against Harmful Inferential Appropriation
Artificial intelligence does not merely collect what persons disclose. It infers what they did not disclose. Statistical systems can derive intimate traits, preferences, risks, and likely future behavior from combinations of apparently mundane data. This creates inferential appropriation, the use of a person’s traces to produce socially consequential claims about that person without meaningful participation in the process. The moral injury may arise even when each individual datum was lawfully obtained, because the invasive act occurs through analytics rather than direct access (Munch, 2021; Mainz, 2024; Rumbold & Wilson, 2019).
Not every inference is wrongful. Ordinary human interaction necessarily includes interpretation and prediction, and many low-stakes inferences are benign. The morally relevant criteria are intimacy, accuracy, stakes, contestability, asymmetry, and downstream consequence. An inference becomes especially objectionable when it attributes stigmatizing traits, determines access to opportunities, operates opaquely, or cannot be challenged by the person whose identity it helps construct. Predictive systems become instruments of domination when they transform probabilistic classifications into durable social facts (Wang et al., 2024; Matar & Murray, 2025).
The Right to Identity Integrity and Truthful Attribution
Identity integrity concerns protection against false attribution, manipulative re-presentation, and substituted agency. It is narrower than a right to control reputation and broader than a right against defamation. Others remain free to form judgments, criticize conduct, and discuss public facts. They do not thereby acquire a moral license to manufacture evidence, simulate speech, or assign conduct to a person through synthetic media. The core wrong is the corruption of the relationship between the person and the person’s socially recognized acts (de Ruiter, 2021; Patella-Rey, 2018).
A dignity-based account explains why the harm persists even when economic damage is absent. Unauthorized synthetic pornography, humiliating replicas, and fabricated political or professional speech treat the represented person as a manipulable object. The person’s face or voice becomes a reusable interface through which others can speak. Such practices do not merely alter reputation; they displace the person’s communicative agency and weaken the evidentiary environment on which social trust depends (Lundgren, 2026; Abousud, 2026).
The Right to Contest and Resist Identity Capture
Rights that cannot be exercised under conditions of asymmetric power risk becoming formal abstractions. Individuals therefore possess a prima facie moral right to resist unauthorized identity capture through legal, social, physical, and technical means. Adversarial privacy tools make this principle concrete. Fawkes, for example, alters images through subtle perturbations designed to interfere with unauthorized facial-recognition training, while adversarial eyeglass patterns demonstrate that physical presentation can be modified to evade automated classification (Shan et al., 2020; Sharif et al., 2016).
Adversarial clothing and image cloaking should not be classified automatically as deception or obstruction. Their moral status depends on the legitimacy of the system being resisted. When surveillance is mass, opaque, unauthorized, or disproportionate, obfuscation can constitute defensive self-presentation, comparable to closing a curtain or declining to answer an intrusive question. The right is not absolute. Legitimate and narrowly tailored identification in grave-risk circumstances may override it. The presumption in favor of resistance is strongest where individuals lack meaningful notice, consent, correction, or withdrawal (Hasan et al., 2023; Smith & Miller, 2021; Véliz, 2022).
The Right to an Open Future and Unchilled Identity Formation
Narrative identity depends on the possibility of revision. Persons develop through experimentation, error, forgetting, forgiveness, and changes of commitment. Permanent searchability and predictive profiling can narrow this horizon by fastening the individual to a past record or probabilistic future. The right to an open future is therefore not a right to erase every true fact. It is a claim that institutions should not convert historical traces into an inescapable identity that preemptively governs future opportunities (Pressly, 2024; Matar & Murray, 2025).
Pervasive observation can also chill identity formation before any formal sanction occurs. A person who reasonably expects continuous recording and retrospective analysis may avoid lawful association, political dissent, unconventional expression, or exploratory conduct. These effects are difficult to measure, which makes them easy to discount in utilitarian calculation, but their diffuse character does not make them morally insignificant. The social conditions of autonomy include spaces in which persons can act without every experiment becoming a permanent institutional record (Macnish, 2018; Matar & Murray, 2025).
Normative Frameworks and Their Limits
No single ethical theory fully explains identity rights in the artificial intelligence age. A libertarian or self-ownership framework provides strong language for consent, exclusion, unauthorized training, and defensive resistance, but it can overstate alienability and understate the relational formation of identity. Kantian ethics explains why synthetic replacement and manipulative profiling are wrong even when no market loss occurs, because they instrumentalize persons and substitute external purposes for autonomous agency. Its weakness is abstraction unless dignity is connected to institutional power.
Utilitarian reasoning is indispensable when surveillance or prediction is defended by public safety, fraud prevention, or aggregate welfare. It requires comparison of benefits and harms, but it can normalize chronic surveillance when chilling effects, unequal exposure, and losses of dignity are difficult to quantify. Rawlsian fairness focuses attention on opacity, arbitrary disadvantage, unequal vulnerability, and the absence of contestability. An ethics of care reveals dependency, coercion, gendered harms, and the relational conditions through which identity is sustained. Virtue ethics asks what habits of honesty, restraint, civic trust, and respect are expressed by institutions that normalize synthetic manipulation or perpetual observation.
The strongest position is therefore pluralistic. Identity governance should combine autonomy-based constraints, dignity protections, fairness requirements, proportionality analysis, attention to relational vulnerability, and institutional virtues of restraint and accountability. The theories converge on a practical conclusion: technological capability does not itself supply moral authorization.
Legal Approaches as Indicators of Moral Priority
Comparative legal approaches reveal differing moral emphases without settling the underlying ethical question. The United States relies heavily on state publicity rights, privacy torts, defamation, sector-specific rules, and constitutional speech protections. This structure recognizes commercial appropriation and reputational injury but provides no unified account of identity as a whole. The European approach gives greater weight to dignity, data protection, biometric sensitivity, and informational self-determination, although data-protection law can still reduce embodied identity to manageable content. United Kingdom scholarship increasingly argues that existing intellectual-property and common-law doctrines underprotect synthetic identity, while Indian courts and comparative scholarship have emphasized consent, voice, likeness, and dignitary protection in the absence of a single publicity statute (Chopra et al., 2025; Hamuľák et al., 2025; Bosher, 2026).
These differences expose a deeper normative choice. A property-centered model supplies remedies and recognizable control mechanisms, but it risks making identity alienable to whoever can purchase rights. A personhood-centered model better explains why some uses remain objectionable even after formal consent, especially where consent is coerced, uninformed, or structurally unequal. The better legal and ethical architecture would preserve limited property-like controls while grounding the core right in dignity, agency, and non-domination.
A Diagnostic Framework for Identity Governance
A disciplined assessment should begin by identifying the stage of the identity pipeline. Is the system observing, identifying, inferring, synthesizing, manipulating, or making a downstream decision? The analysis should then ask whether the practice destroys contextual obscurity, appropriates intimate information by inference, substitutes fabricated agency for real agency, locks the person to a searchable past, or chills experimentation and dissent.
The next inquiry concerns justification. What specific interest is advanced, what evidence shows that the practice is necessary, and are less intrusive alternatives available? A generalized appeal to innovation, convenience, or security is insufficient. Proportionality requires that the expected benefit be commensurate with the severity, duration, and distribution of identity harms. Fairness requires attention to who is observed, who controls the system, who bears false positives, and who can obtain correction.
Finally, a morally defensible system must preserve agency after deployment. Persons should have meaningful opportunities to know that identity processing is occurring, contest classifications, correct false information, restrict downstream use, resist unjustified capture, and obtain remedies for synthetic replacement or manipulated attribution. These requirements do not eliminate all surveillance, prediction, or synthetic media. They establish the conditions under which such practices might be justified.
Conclusion
The artificial intelligence age requires a shift from narrow data protection to identity governance. Persons do not possess an absolute ownership right over every representation or judgment involving them, but they do possess strong, plural, and qualified moral rights over the processes that create, observe, predict, and manipulate socially consequential identities. These rights include protection against unauthorized synthetic replacement, continuous biometric identification that destroys practical obscurity, harmful inferential appropriation, false attribution, substituted agency, and institutional practices that close the future by transforming the past into a permanent predictive sentence.
The concept of digital integrity supplies the unifying principle. Digital manifestations should be protected when they function as extensions of embodied agency, social recognizability, or narrative self-authorship. Informational self-determination supplies the procedural requirement that persons retain meaningful power over how those manifestations are formed and used. Neither principle is absolute. Expression, public memory, journalism, art, and legitimate public-safety needs impose real limits. However, those limits must be demonstrated through specific, proportionate, and accountable justification.
The central moral error is to treat identity as unrestricted raw material merely because technology makes extraction, inference, and fabrication inexpensive. A person is not reducible to a vector, a risk score, a synthetic likeness, or a searchable archive. Ethical governance begins when institutions recognize that identity is not only information about a person. It is one of the primary media through which the person acts, is recognized, and remains capable of becoming someone not yet predicted.
Bibliography
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: A layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence, 9, 1873975. https://doi.org/10.3389/frai.2026.1873975
Alexander, H., Simon, J., & Pinard, F. (2026). How Should the Law Treat Future AI Systems? Fictional Legal Personhood versus Legal Identity. Journal of Law, Technology, & the Internet, 17(1), Article 3. https://scholarlycommons.law.case.edu/jolti/vol17/iss1/3
AMA Journal of Ethics. (2008). Personhood and Autonomy in Multicultural Health Care Settings. AMA Journal of Ethics. https://journalofethics.ama-assn.org/article/personhood-and-autonomy-multicultural-health-care-settings/2008-03
Andrenacci, G. (2026, May). What Anthropic, OpenAI, and Google reveal — and conceal — when they try to encode values into machines. TechnoSophia |
Medium. https://medium.com/tecnosophia/what-anthropic-openai-and-google-reveal-and-conceal-when-they-try-to-encode-values-into-208d9d53c3ff
Anthropic. (n.d.). Claude’s Constitution. https://www.anthropic.com/constitution
Atkinson, M. (2019). The Right of Publicity for North Carolina: Focused on the Identity-Holder, Privacy-Based and Limited Alienability. North Carolina Central University Science & Intellectual Property Law Review, 12(1), Article https://archives.law.nccu.edu/siplr/vol12/iss1/3
Bartz-Beielstein, T. (2019). Why We Need an AI-Resilient Society. arXiv. https://arxiv.org/abs/1912.08786
Bosher, H. (2026). Dr Hayleigh Bosher—written evidence (AIC0025). UK Parliament Committees.
Caversham House. (n.d.). Library. https://cavershamhouse.com/library.html
Chopra, P., Sony, R. A. L., & Chopra, S. (2025). Generative AI, Copyright and Personality Rights: A Comparative Legal Perspective. Legal Issues in the Digital Age, 6(3), 23-51. https://doi.org/10.17323/2713-2749.2025.3.23.51
Common Sense Media. (2026, July 14). Google Search: AI Overview & AI Mode. Youth AI Safety Institute. https://institute.commonsensemedia.org/risk-assessments/google-search
Craig, C., & Dhonchak, A. (2022). Against Integrity: A Feminist Theory of Moral Rights, Creative Agency & Attribution. Osgoode Digital
Commons. https://digitalcommons.osgoode.yorku.ca/cgi/viewcontent.cgi? article=1364&context=all_papers
de Ruiter, A. (2021). The Distinct Wrong of Deepfakes. Philosophy & Technology, 34, 1311-1332. https://doi.org/10.1007/s13347-021-00459-2
Degges-White, S. (2021, October). Personal and Social Identity: Who Are You Through Others’ Eyes. Psychology Today. https://www.psychologytoday.com/us/blog/lifetime-connections/202110/personal-and-social-identity-who-are-you-through-others-eyes
Digital Trends. (n.d.). Anthropic research says AI can mass expose of anonymous internet accounts. https://www.digitaltrends.com/social-media/anthropic-research-says-ai-can-mass-expose-of-anonymous-internet-accounts/
Dogan, S. L., & Lemley, M. A. (2006). What the Right of Publicity Can Learn from Trademark Law. Stanford Law Review, 58, 1161, https://scholarship.law.bu.edu/faculty_scholarship/420
Dunn, J. E. (2026, January 22). Anthropic’s Claude AI gets a new constitution embedding safety and ethics. CIO. https://www.cio.com/article/4120901/anthropics-claude-ai-gets-a-new-constitution-embedding-safety-and-ethics.html
Efat, S. I. J. (2026). Personality Rights, Detection Bias and Effective Remedies for AI-Generated Identity Harm: A Comparative Study of India and Bangladesh. International Journal of Law and Policy, 11(1), 1-25. https://doi.org/10.47604/ijlp.3876
Elad, B. (2026, April 30). Top AI Fraud Prevention Platforms That Evolve With Threats 2026. SQ Magazine. https://sqmagazine.co.uk/ai-fraud-prevention-threats/
Fox, Y. (2026). Using Generative AI in Historical Practice. Elements in Historical Theory and Practice. Cambridge University Press. https://doi.org/10.1017/9781009740067
Google. (2018, June 7). AI at Google: our principles. https://blog.google/innovation-and-ai/products/ai-principles/
Google. (n.d.). AI Principles. Google AI. https://ai.google/principles/
Hamuľák, O., Vardanyan, L., & Kocharyan, H. (2025). From Cells to Codes: The Legal Transformation of Digital Identity, Integrity, and Self-Determination in the Digital Age. Acta Universitatis Lodziensis. Folia Iuridica, 113, 129-142. https://doi.org/10.18778/0208-6069.113.07
Hurshman, C., Voinea, C., D’Alessandro, W., et al. (2025). Beyond deepfakes: ‘digital duplicates’ or AI simulations of real people – an international ethical consensus. Pre-print.
Insight Partners. (n.d.). The jagged frontier of generative AI: A conversation with Ethan Mollick. https://www.insightpartners.com/ideas/generative-ai-ethan-mollick/
Insight Partners. (n.d.). What Leaders Get Wrong About AI Adoption, According to Ethan Mollick. https://www.insight.com/en_US/content-and-resources/insight-on/inside-the-new-reality-of-ai-at-work-with-ethan-mollick.html
Interdisciplinary Encyclopedia of Religion and Science. (n.d.). Self, Narrative Perspective of. https://inters.org/self-narrative-perspective
International Trademark Association. (n.d.). Right of Publicity. https://www.inta.org/topics/right-of-publicity/
Journal of Accountancy. (2026, July). A guide to fighting AI-fueled AP/AR fraud. https://www.journalofaccountancy.com/issues/2026/jul/a-guide-to-fighting-ai-fueled-ap-ar-fraud/
Lundgren, B. (2026). Can Deepfakes Violate an Individual’s Moral Right to Privacy?. Ethical Theory and Moral Practice, 29(1), 125-139. https://doi.org/10.1007/s10677-025-10514-y
Lyon, D. (2006). Surveillance, Power and Everyday Life. In Oxford Handbook of Information and Communication Technologies. https://www.antoniocasella.eu/nume/Lyon_oxford_handbook_2009.pdf
Lyon, D. (2007). Surveillance Studies: An Overview. Polity Press. https://www.researchgate.net/publication/26526923_David_Lyon_Surveillance_ Studies_An_Overview
Mercadal, T. (2021). Identity formation. EBSCO Research Starters. https://www.ebsco.com/research-starters/psychology/identity-formation
Munch, L. A., & Mainz, J. T. (2023). The Moral Significance of Privacy Dependencies. Philosophy & Technology. https://doi.org/10.1007/s13347-023-00680-1
National Library of Medicine. (2018, May 17). Patron Guide to Copyright and Historical Materials – Privacy and Publicity
Rights. https://www.nlm.nih.gov/hmd/copyright/patron-guide-privacy-and-publicity-rights.html
Olson, E. T. (2023). Personal Identity. Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/identity-personal/
Oxford Academic. (n.d.). Canadian copyright protections for neurodata: ethical and legal implications. Journal of Law and the
Biosciences. https://academic.oup.com/jlb/article/13/1/lsag008/8524171
Oxford Academic. (n.d.). Re-thinking international human rights law’s approach to identity in light of surveillance and AI. Human Rights Law Review. https://academic.oup.com/hrlr/article/25/3/ngaf016/8157328
Oxford Academic. (n.d.). Surveillance and identity: conceptual framework and formal models. Cybersecurity. https://academic.oup.com/cybersecurity/article/3/3/145/4748787
PMC – NIH. (n.d.). Respecting privacy and upholding confidentiality: core ethical duties. https://pmc.ncbi.nlm.nih.gov/articles/PMC12747445/
PMC – NIH. (n.d.). The dynamics of big data and human rights: the case of scientific research. https://pmc.ncbi.nlm.nih.gov/articles/PMC5124070/
PMC – NIH. (n.d.). The interplay of social identity and norm psychology in the evolution of human groups. https://pmc.ncbi.nlm.nih.gov/articles/PMC9869443/
Pressly, L. (2024). The Right to Be Forgotten and the Value of an Open Future. Ethics, 135(1). https://doi.org/10.1086/731431
Proof. (2026). Agentic AI Needs Verifiable Records to Be Trusted. https://www.proof.com/blog/agentic-ai-needs-verifiable-records
Shaheen, U., & Naz, N. (2025). Reevaluating Patient Privacy and Autonomy: Cross-Cultural and Legislative Perspectives. medtigo J Med., 3(3), e3062335. https://doi.org/10.63096/medtigo3062335
Shan, S., Wenger, E., Zhang, J., Li, H., Zheng, H., & Zhao, B. Y. (2020). Fawkes: Protecting Privacy against Unauthorized Deep Learning Models. In Proceedings of the 29th USENIX Security Symposium. https://www.usenix.org/conference/usenixsecurity20/presentation/shan
Sharif, M., Bhagavatula, S., Bauer, L., & Reiter, M. K. (2016). Accessorize to a Crime: Real and Stealthy Attacks on State-of-the-Art Face Recognition. In Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security (CCS ’16). ACM.
Taylor, J., & Krishna, K. (2026, April). Context-maxxing: A path to cognitive agency with generative AI. Working Paper 196. Brookings
The Artificial Self: Characterising the landscape of AI identity. (2026). arXiv. https://arxiv.org/html/2603.11353v1
Towards a Human Right to Psychological Continuity? Reflections on the Rights to Personal Identity, Self-Determination, and Personal Integrity. (n.d.). Brill. https://brill.com/view/journals/eclr/5/2/article-p199_005.xml
University of Chicago News. (2026). Eight UChicago scholars awarded prestigious Sloan Fellowships in 2026. https://news.uchicago.edu/story/eight-uchicago-scholars-awarded-prestigious-sloan-fellowships-2026
Unknown Author. (n.d.). Beyond Human Rights. https://mau.diva-portal.org/smash/get/diva2:1966808/FULLTEXT02.pdf
Unknown Author. (n.d.). Moral Rights Basics. https://cyber.harvard.edu/property/library/moralprimer.html
Unknown Author. (n.d.). Self and Identity – An Introduction to Social
Psychology. https://opened.tesu.edu/introsocialpsychology/chapter/self-and-identity/ Unknown Author. (n.d.). The ethical application of biometric facial recognition technology.
Véliz, C. (2022). Privacy in Social Media. Journal of Practical Ethics, 9(2).
Véliz, C., & Law, S. (2023). Artificial Intelligence and Data Harvesting: An Interview with Carissa Véliz. Think, 22, 59-62. https://doi.org/10.1017/S1477175622000215
VerifyWise. (n.d.). Google AI Principles. VerifyWise AI Governance Library. https://verifywise.ai/ai-governance-library/policies-and-internal-governance/google-ai-principles
Wang, X., Wu, Y. C., Zhou, M., & Fu, H. (2024). Beyond surveillance: privacy, ethics, and regulations in face recognition technology. Frontiers in Big Data, 7, 37465. https://doi.org/10.3389/fdata.2024.1337465
Weisbord, A. (2016). A Copyright Right of Publicity. Fordham Law Review, 84(6), 2804-2844. https://fordhamlawreview.org/wp-content/uploads/2016/04/Weisbord_May.pdf
Wenger, E. (2020). Fawkes: Protecting Personal Privacy against Unauthorized Deep Learning Models (USENIX Security 2020) [Video]. YouTube.
Wickert, C. (2025, July 29). David Lyon – Surveillance Studies: An Overview (2007). SozTheo. https://soztheo.com/criminology/key-works-in-criminology/david-lyon-surveillance-studies-an-overview-2007/
Zimmer, M. (2015). Research Ethics in the Big Data Era: Addressing Conceptual Gaps for Researchers and IRBs. Big Data Ethics. https://bigdata.fpf.org/wp-content/uploads/2015/12/Zimmer-Research-Ethics-in-the-Big-Data-Era.pdf