Skip Navigation
Skip to contents

JPMPH : Journal of Preventive Medicine and Public Health

OPEN ACCESS
SEARCH
Search

Articles

Page Path
HOME > J Prev Med Public Health > Volume 59(4); 2026 > Article
Perspective
Protecting Informational Self-determination in AI-driven Precision Oncology: Privacy, Ethics, and Governance Challenges
Minsoo Jung1,2corresp_iconorcid
Journal of Preventive Medicine and Public Health 2026;59(4):428-431.
DOI: https://doi.org/10.3961/jpmph.26.094
Published online: April 27, 2026
  • 1,241 Views
  • 57 Download

1Department of Health Science, Dongduk Women’s University, Seoul, Korea

2Center for Community-Based Research, Dana-Farber Cancer Institute, Boston, MA, USA

Corresponding author: Minsoo Jung, Department of Health Science, Dongduk Women’s University, 60 Hwarang-ro 13-gil, Seongbuk-gu, Seoul 02748, Korea, E-mail: mj748@dongduk.ac.kr
• Received: January 29, 2026   • Revised: March 4, 2026   • Accepted: March 26, 2026

Copyright © 2026 The Korean Society for Preventive Medicine

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

prev next
  • Artificial intelligence (AI) is rapidly transforming cancer care, from diagnosis and risk prediction to treatment planning and drug development. These advances depend on the large-scale integration of genomic, imaging, and clinical data, intensifying privacy and ethical concerns. This study examines emerging challenges in AI-driven precision oncology and explores strategies for protecting patients’ informational self-determination through a narrative ethical and policy analysis informed by recent empirical and technical research. Even data considered de-identified, including molecular and imaging profiles, may enable inference of sensitive attributes through machine-learning analytics. Although many patients support the clinical promise of AI, they express concerns about secondary data use without meaningful consent, commercialization, and security breaches. Privacy violations threaten autonomy, confidentiality, and trust and may exacerbate existing inequities. Current regulatory frameworks, including the General Data Protection Regulation and the Health Insurance Portability and Accountability Act, focus primarily on identifiability and the removal of explicit identifiers. However, AI systems generate risks not only through re-identification but also through inferential analytics that derive new privacy-relevant information, such as disease susceptibility or familial risk, from ostensibly anonymized data. Addressing these gaps requires multilayered governance that combines privacy-preserving technologies, dynamic consent, data minimization, transparency, and strengthened legal safeguards to sustain public trust.
Artificial intelligence (AI) is transforming cancer care across detection, diagnosis, prognosis, and treatment optimization. From imaging-based tumor detection to genomic profiling and therapeutic selection, AI systems depend on the large-scale integration of molecular, imaging, and clinical data [1,2]. These advances promise improved accuracy and personalization. However, the same data-intensive infrastructure raises substantial ethical and privacy concerns. AI systems can derive predictive and inferential insights from sensitive biological information, creating risks to autonomy, confidentiality, and trust [2–4].
Central to these concerns is informational self-determination: the capacity of individuals to understand and influence how their health information is collected, linked, analyzed, and reused. Unlike traditional informed consent, which authorizes discrete interventions, informational self-determination extends across the data lifecycle. It involves temporal continuity beyond 1-time authorization, control over downstream analytics, and recognition of future algorithmic uses that may be unpredictable (Table 1). In AI-driven precision oncology, where data circulate persistently and new inferences emerge over time, episodic consent provides only limited protection [4–6].
This study examines emerging privacy and ethical challenges in AI-based oncology and identifies governance strategies for safeguarding informational self-determination in data-intensive medicine.
AI technologies are now embedded across the cancer care continuum. Machine-learning approaches support tumor detection, histopathological classification, treatment-response prediction, and drug discovery [2,7]. Precision oncology integrates genomic sequencing, imaging, clinical histories, and, in some cases, environmental data into interoperable analytic platforms.
Cancer-related information has therefore shifted from discrete records to a longitudinal, reusable data resource. These data are frequently repurposed for secondary research, algorithm refinement, and commercial development [4,8]. This cumulative reuse increases exposure and diminishes meaningful patient oversight. As a result, AI-driven oncology operates as a dynamic data ecosystem that challenges traditional consent-based models of control.
Informational self-determination encompasses awareness of data practices, the ability to authorize or refuse specific uses, ongoing oversight of secondary applications, and protection against identity linkage or predictive inference. In oncology, these dimensions are particularly salient because health data may reveal genetic susceptibility, prognostic probabilities, and familial implications.
AI challenges each of these dimensions. Data uses evolve beyond their original purposes; algorithms infer attributes that were never explicitly disclosed; and patients rarely understand how their information circulates across digital infrastructures [3–6]. Even when consent is obtained, individuals seldom revisit these decisions as analytic capabilities evolve. Consequently, control becomes largely formal rather than substantive.
These developments suggest that the challenge extends beyond improving consent procedures to rethinking governance structures capable of mediating competing rights and public interests.
The erosion of informational self-determination has implications for core ethical principles. Autonomy is weakened when patients cannot meaningfully influence AI-driven data uses [4,9]. Confidentiality is destabilized as re-identification and inferential analytics undermine assumptions about anonymization [5,10]. Justice is threatened when privacy risks and algorithmic harms disproportionately affect marginalized populations [8]. Trust—an essential component of oncology care—may decline when patients perceive commercialization or misuse of sensitive data [8,9].
These harms arise not only from isolated breaches but also from the structural features of AI ecosystems, including scale, persistence, opacity, and commercial integration. Marginalized populations are often represented in fragmented or biased datasets while lacking the resources to contest harmful uses. Biased training data may produce less accurate predictions or treatment recommendations, thereby reinforcing disparities. Accordingly, justice extends beyond algorithmic fairness to include the equitable distribution of both the benefits and burdens of AI-enabled oncology.
Cancer-related data amplify privacy risks. Genomic information is biologically persistent and difficult to anonymize [5,10]. Diagnoses may carry stigma and potential employment consequences, while prognostic modeling generates sensitive predictions that extend beyond immediate clinical care.
Oncology data also implicate family members through shared genetic risk. AI systems may reveal information about relatives who have not provided consent. Genomic inference can generate privacy-relevant insights about biologically related individuals, challenging purely individual models of consent. Cancer data are therefore ethically distinctive and warrant heightened governance protections [4,8].
Traditional privacy protections emphasize de-identification and pseudonymization. However, AI systems can reconstruct identities or infer sensitive attributes from ostensibly anonymized datasets through linkage attacks or model inversion [5,10]. Genomic sequences function as near-unique identifiers.
Privacy-preserving approaches, including federated learning and differential privacy, can reduce centralized data exposure [10]. However, in precision oncology, predictive accuracy directly influences treatment selection and survival estimation. Even modest reductions in model performance may have clinically meaningful consequences. These trade-offs are therefore both normative and technical. Determining acceptable levels of uncertainty requires balancing clinical benefit against privacy protection. Regulatory frameworks that classify data as either identified or de-identified inadequately address evolving inferential capabilities.
AI development in oncology frequently involves partnerships between health care institutions and technology companies. In these arrangements, patient data may be transformed into proprietary assets used to train commercially valuable algorithms [4,8]. Patients typically have limited visibility into the contractual terms governing these collaborations or into mechanisms for benefit sharing.
This asymmetry in knowledge and control weakens meaningful patient autonomy and raises concerns about exploitation and distributive justice. When commercial imperatives dominate governance structures, informational self-determination may be further constrained.
Empirical evidence indicates that many patients with cancer support the use of AI in clinical care but express concerns regarding secondary data use, commercialization, and data misuse in the absence of meaningful consent [3,9]. Transparency and accountability are consistently identified as essential conditions for maintaining trust.
Trust is ethically foundational in oncology, where patients disclose highly sensitive information under conditions of vulnerability. Erosion of trust may reduce willingness to participate in research and data sharing, thereby impeding scientific progress.
Safeguarding informational self-determination requires coordinated reforms: (1) Deployment of privacy-preserving technologies that minimize exposure while enabling inter-institutional learning [10]; (2) Dynamic and granular consent mechanisms allowing patients to revisit permissions during care transitions [4]; (3) Data minimization and purpose limitation embedded in dataset curation and model training; (4) Transparency and explainability requirements supporting accountability [4,8]; (5) Strengthened legal protections against re-identification and discriminatory misuse; (6) Equitable benefit-sharing mechanisms in commercial partnerships; (7) Explicit recognition of relational and familial privacy, ensuring that 1 patient’s consent does not automatically override the privacy interests of relatives.
No single intervention is sufficient. Effective protection depends on integrated ethical, technical, and legal governance frameworks that are adaptable across regulatory contexts.
AI is now embedded within modern cancer care infrastructures. Systems that enhance detection and treatment depend on continuous flows of genomic, imaging, and clinical data that extend beyond individual clinical encounters. While enabling precision oncology, this transformation places strain on established norms of privacy and patient control. Consent models designed for discrete uses are poorly aligned with iterative machine-learning processes and cross-sector data sharing, leaving patients with limited ongoing agency.
Effective responses must extend beyond procedural consent to structural governance. Transparency, limitations on secondary data use, accountability for both clinical and commercial actors, and equity-focused oversight are essential. Vulnerable populations face disproportionate risks from biased data and inadequate governance, which may amplify existing disparities.
These challenges are particularly pronounced in health systems with centralized data infrastructures. In the Korea, the single-payer National Health Insurance Service enables large-scale linkage of insurance claims, cancer registry data, and hospital records, creating substantial opportunities for AI-driven oncology while also increasing the risk of cross-dataset inference. National-level transparency and accountability are therefore essential. In cultural contexts where family involvement in cancer decision-making is common, genomic data carry relational implications that extend beyond individual patients.
Informational self-determination should therefore be understood not only as an individual right but also as a public health framework for negotiating privacy, familial interests, and collective data infrastructures. Protecting it—beyond the procedural adequacy of consent—is central to sustaining public trust in data-driven cancer care.

Ethics Statement

This study is a perspective article based on secondary sources and does not require ethical approval.

Conflict of Interest

The author has no conflicts of interest associated with the material presented in this paper.

Funding

None.

Acknowledgements

None.

Author Contributions

All work was done by Jung M.

Table 1
Conceptual distinction between consent and informational self-determination in AI-driven oncology
Dimension Traditional informed consent Informational self-determination
Temporal scope One-time authorization Ongoing governance across the data lifecycle
Object of control Specific intervention or study participation Data linkage, reuse, analytics, and inference
Risk orientation Foreseeable and disclosed uses Future, secondary, and inferential uses
Ethical focus Procedural validity of decision-making Sustained autonomy and structural control

AI, artificial intelligence.

Figure & Data

References

    Citations

    Citations to this article as recorded by  

      Protecting Informational Self-determination in AI-driven Precision Oncology: Privacy, Ethics, and Governance Challenges
      Protecting Informational Self-determination in AI-driven Precision Oncology: Privacy, Ethics, and Governance Challenges
      Dimension Traditional informed consent Informational self-determination
      Temporal scope One-time authorization Ongoing governance across the data lifecycle
      Object of control Specific intervention or study participation Data linkage, reuse, analytics, and inference
      Risk orientation Foreseeable and disclosed uses Future, secondary, and inferential uses
      Ethical focus Procedural validity of decision-making Sustained autonomy and structural control
      Table 1 Conceptual distinction between consent and informational self-determination in AI-driven oncology

      AI, artificial intelligence.


      JPMPH : Journal of Preventive Medicine and Public Health
      TOP