AUTOMATING END-OF-LIFE DECISIONS? TECHNOLOGY, EUTHANASIA, AND THE FRAGILITY OF LEGAL SAFEGUARDS

AUTOMATING END-OF-LIFE DECISIONS? TECHNOLOGY, EUTHANASIA, AND THE FRAGILITY OF LEGAL SAFEGUARDS

Abstract: Euthanasia remains one of the most widely debated issues globally. While proponents emphasize bodily autonomy and the right to make end-of-life decisions, opponents rely on the sanctity of life to oppose its legalization. In response to these competing perspectives, certain jurisdictions have legalized euthanasia and assisted suicide under defined regulatory frameworks intended to prevent abuse and ensure compliance with established safeguards. However, in practice, regulatory enforcement has been inconsistent, resulting in significant compliance gaps in some jurisdictions. The increasing integration of technology into healthcare further complicates this regulatory landscape. Although technological advancements have generally improved medical practice, their application in the context of euthanasia and assisted suicide raises new legal and ethical concerns, potentially weakening existing safeguards. This paper does not engage in the debate over the legalization of euthanasia but instead examines the impact of technological integration on its regulation and practice. It argues that technological developments intensify regulatory challenges and calls for stricter and technologically responsive regulatory mechanisms to mitigate foreseeable risks.

Keywords: Euthanasia, Assisted Suicide, Technology, Regulation, Artificial Intelligence (AI)

INTRODUCTION

The concept of euthanasia has remained controversial since its emergence in medicolegal discourse. Derived from the Greek words eu (good) and thanatos (death), it literally translates to "good death." Broadly, euthanasia refers to the intentional hastening of a patient's death to relieve suffering. Active euthanasia involves a deliberate act, typically the administration of lethal substances to end the life of a terminally or incurably ill patient, whereas passive euthanasia entails the withholding or withdrawal of life-sustaining treatment. In relation to consent, active euthanasia may be voluntary, non-voluntary, or involuntary1. Assisted suicide, by contrast, occurs where an individual intentionally ends his or her own life with the knowledge or means provided by another2. While a technical distinction exists -- the patient performs the final act in assisted suicide -- both practices are often collectively described under the broader concept of assisted dying3.

As of 2025, both voluntary active euthanasia and physician-assisted suicide are legally permitted under regulated frameworks in several jurisdictions. Countries that allow voluntary euthanasia include Belgium, Canada, Colombia, Ecuador, Luxembourg, the Netherlands, New Zealand, Spain, and Uruguay, along with all Australian states under voluntary assisted dying schemes. Assisted suicide is also permitted in jurisdictions such as Switzerland and several states in the US. In some countries, legislation has been enacted but is pending full implementation4.

Despite legalization, regulatory implementation remains complex. Scholarship indicates that eligibility criteria have evolved in certain jurisdictions, at times expanding beyond their initial scope. Moreover, persistent challenges arise in assessing voluntariness, coercion, and decisional capacity, particularly among vulnerable populations. Variations in reporting compliance and oversight further demonstrate that formal safeguards do not always translate into uniform regulatory practice5. Thus, even within conventional medical frameworks, the governance of assisted dying remains delicate and contested.

Against this already fragile regulatory background, the growing integration of technology into medical decision-making introduces a further layer of complexity. Artificial intelligence is now routinely employed in diagnostics, predictive modelling, personalised medicine, and robotic surgery. Its extension into end-of-life decision-making -- through algorithmic eligibility assessments, automated drug-delivery systems, or technologically mediated access platforms -- raises profound ethical and legal questions. In this context, technology is no longer merely a passive instrument but risks becoming an active structural component in the facilitation of assisted dying. Regulatory systems originally designed to supervise human clinical judgment may therefore prove inadequate when confronted with increasingly automated or technologically embedded processes6.

This paper examines the emerging technological input in euthanasia and assisted dying, analysing the foreseeable risks associated with mechanisation and automation in end-of-life decision-making. It argues that without deliberate and anticipatory regulatory intervention, the integration of such technologies may outpace existing safeguards, thereby intensifying vulnerabilities within an already sensitive legal domain.

REGULATORY FRAMEWORKS AND EXISTING CHALLENGES

The regulation of euthanasia and assisted suicide raises complex legal, medical, ethical, and political questions. Although legalization in certain jurisdictions has been accompanied by statutory safeguards, challenges persist in translating these safeguards into consistent regulatory practice. In particular, the assessment of decision-making capacity is not always conducted with sufficient scrutiny, and weaknesses in oversight mechanisms undermine effective enforcement. Furthermore, in contexts of rising healthcare costs, concerns have emerged that assisted dying may risk being misapplied as a cost-efficient alternative rather than a strictly autonomy-based medical decision7.

The regulatory challenges associated with legalized euthanasia and assisted suicide have been critically examined by Jose Pereira8, whose analysis forms the foundation for much of the discussion in this section. Pereira argues that statutory safeguards may create an impression of effective control while failing to ensure consistent enforcement in practice. His work demonstrates how formal regulatory mechanisms, although detailed in legislation, often encounter significant limitations during implementation.

The legalization of euthanasia and assisted suicide has typically been accompanied by detailed statutory safeguards designed to protect vulnerable populations and ensure that only eligible individuals receive support in ending their lives. These safeguards often include requirements for explicit patient consent, documentation of unbearable suffering, mental capacity assessments, multiple physician approvals, and mandatory reporting to oversight bodies. However, empirical evidence from jurisdictions with long-standing assisted dying regimes reveals persistent and significant challenges in translating these legal safeguards into practice.

Building on regulatory experiences in jurisdictions such as the Netherlands and Belgium, Pereira's analysis illustrates that legal safeguards may be interpreted expansively over time.

Although legislation imposes strict eligibility criteria, key concepts such as "unbearable suffering" have gradually been interpreted more broadly, resulting in applications that extend beyond the original legislative intent. This regulatory expansion raises concerns about the practical containment of assisted dying within its prescribed legal boundaries.

Empirical data further reveal weaknesses in reporting and oversight mechanisms. Physicians often report cases based on self-assessment, with limited independent verification, and enforcement for non-compliance remains rare. In certain jurisdictions, studies have indicated that a significant proportion of cases have not been formally reported despite mandatory obligations, undermining transparency and accountability. Instances of assisted deaths conducted without clearly documented consent further illustrate the gap between statutory requirements and practical implementation.

These implementation challenges demonstrate that formal regulatory frameworks do not automatically ensure effective oversight or protection. The persistent gap between legal safeguards and clinical practice reflects structural fragility within existing systems. This regulatory weakness provides the backdrop against which emerging technological mediation must be examined, as automation and technological integration may further complicate already strained oversight mechanisms.

TECHNOLOGICAL INPUT INTO EUTHANASIA AND ASSISTED SUICIDE

Technological intervention in assisted dying has evolved from a supportive tool within medical practice to a mechanism that, in certain contexts, directly structures and facilitates access to death outside conventional clinical frameworks. In his sociotechnical analysis of euthanasia activism, Brian Martin9 highlights the case of Australian physician and advocate Dr. Philip Nitschke and the organisation Exit International, which sought technological routes to enable self-directed death outside conventional legal pathways. Martin notes that Nitschke's efforts included the development and dissemination of "do-it-yourself" technological means for dying with dignity, reflecting a shift from legalisation as the central strategy to technology as an instrument of access and empowerment. This form of technological facilitation illustrates how innovation can intersect with assisted dying in ways that challenge existing regulatory assumptions.

However, technological influence in assisted dying extends beyond advocacy into the development of physical mechanisms engineered to administer or facilitate death.

Historically, technological mediation in this field is not entirely novel. Jack Kevorkian -- widely known as "Dr. Death" -- assisted over 130 individuals in ending their lives through his self-designed "mercy machine." The device consisted of two primary systems: the "Thanatron" and the "Mercitron." The Thanatron operated through an intravenous mechanism in which the patient activated the system to release sequential substances -- beginning with thiopental to induce unconsciousness, followed by a lethal dose of potassium chloride. The Mercitron, reportedly used more frequently, functioned through the release of a lethal gas delivered via a mask. Importantly, both systems were structured so that the patient performed the final activating act, embedding technological intervention within a framework that emphasised asserted autonomy10.

Building on this mechanical foundation, later developments reflected a more systematised and computer-assisted approach. The Deliverance Machine, developed by Philip Nitschke, incorporated programmed prompts requiring users to confirm their intent before a lethal dose could be administered. Unlike purely mechanical systems, it integrated procedural verification into its operational logic. As described by Exit International, the device demonstrates how technological architecture can structure not only the delivery of death but also the process of consent confirmation through automation11.

Technological expression in this domain also extends beyond functional devices to symbolic installations. "My Beautiful Chair," documented by Exit International, represents an artefact that reflects the intersection of autonomy, technology, and assisted dying. Although not a medical device in the conventional sense, the installation functions as both conceptual design and technological representation, framing assisted dying within broader narratives of dignity, bodily agency, and mechanised facilitation12.

The most widely discussed contemporary example is the Sarco Pod -- a 3D-printed capsule designed to induce death through nitrogen hypoxia once activated by the user. According to descriptions provided by Exit International, the Sarco is engineered as a self-contained and self-initiated mechanism that minimises direct physician involvement and replaces it with machine-mediated activation. Its exhibition and continued development mark a shift from experimental prototype to publicly displayed technological proposition, demonstrating how design can materialise assisted dying as an automated and controlled process13.

Beyond the standard model, media reports have described proposed modifications such as the "Double Sarco," a speculative variant intended to allow two individuals to initiate the process simultaneously. The concept reportedly involves dual activation mechanisms requiring coordinated consent prior to operation. Although it has not been implemented or formally developed as a functioning device, its discussion reflects ongoing technological speculation around collective autonomy and mechanised facilitation of death. As such, it raises additional ethical and regulatory questions concerning consent verification, interaction between users, and accountability14.

Speculative technological extensions continue further with conceptual designs such as the Euthanasia Coaster -- a theoretical engineering proposal conceived by Julijonas Urbonas in 2010. The design envisions a roller coaster engineered to induce death through sustained exposure to extreme gravitational forces capable of producing cerebral hypoxia. While never constructed, the concept has been examined in design discourse and featured in the Museum of Modern Art's exhibition on speculative technology and violence, situating it within broader cultural debates about autonomy, engineered mortality, and technological imagination15, 16.

More recently, reports have emerged of a proposed implantable "kill switch" intended for individuals with early-stage dementia. The proposal allegedly involves an implanted device that emits periodic activation signals and requires user deactivation; failure to respond -- potentially due to cognitive decline -- would trigger the release of a lethal substance. The device remains purely speculative and has neither been developed nor approved as a medical technology. Nevertheless, its discussion underscores emerging ethical concerns surrounding automation, consent, and technological control in end-of-life decision-making17.

Taken together, these developments demonstrate a trajectory in which technological mediation in assisted dying has progressed from mechanical assistance to programmed automation and increasingly speculative self-executing systems. The emergence of such technologies challenges regulatory frameworks that were primarily designed around physician oversight and human decision-making, raising questions about how legal systems should respond to mechanised and potentially autonomous forms of assisted death.

TECHNOLOGICAL INTEGRATION AND THE DEEPENING OF REGULATORY GAPS

Much of the existing scholarship addressing AI in euthanasia and assisted dying is framed in ethical terms, focusing on autonomy, bias, human dignity, and moral responsibility. While these analyses are valuable, they also reveal structural weaknesses within regulatory systems. Rather than treating them solely as moral critique, this section draws from them to demonstrate how ethical concerns expose concrete regulatory deficiencies, particularly where legal frameworks remain silent on technological participation in end-of-life decision-making.

Advances in artificial intelligence and automated systems have introduced new dimensions to the governance of euthanasia and assisted suicide. AI systems can analyse large datasets to support clinical assessments and predict disease progression, potentially influencing eligibility determinations18. However, assisted dying statutes were largely designed around human clinical judgment and physician responsibility. The introduction of algorithmic systems therefore raises unresolved questions about permissible delegation, oversight, and accountability; areas that existing regulations rarely address explicitly.

A key regulatory concern relates to misinformation and the legal status of AI-generated medical guidance. AI tools may produce recommendations or informational outputs that influence clinical or patient decisions19, yet many jurisdictions do not clearly classify such systems under medical device regulation or specify standards for their use in assisted dying contexts. This ambiguity creates uncertainty about the legal validity of relying on AI outputs in capacity assessments or procedural evaluations. If inaccurate or biased information is generated, current regulatory mechanisms often lack clear accountability structures to address resulting harm.

Closely connected to this is technological over-reliance and system error. While statutes assume that capacity and voluntariness assessments are conducted through direct professional engagement, increased reliance on AI-driven analysis risks diluting human responsibility. Algorithmic systems are vulnerable to data bias and limited transparency, and they typically do not provide reasoned explanations for their conclusions20. A malfunctioning system could therefore distort eligibility determinations. However, existing frameworks do not clearly define limits on algorithmic reliance nor mandate explicit standards for human supervision when such tools are used in sensitive life-ending decisions.

Autonomy remains central to assisted dying regulation, yet technological mediation complicates its exercise. If algorithmic recommendations significantly shape clinical judgments, decision-making may become indirectly structured by system design rather than purely by patient choice and professional deliberation21. Most statutes predate advanced AI systems and do not specify boundaries for algorithmic participation in capacity evaluation or procedural certification. This legislative silence reflects a regulatory gap concerning the extent to which technology may influence determinations traditionally reserved for human actors.

Technological facilitation also introduces concerns about regulatory expansion -- the slippery slope argument22. Automated systems that streamline eligibility assessment may unintentionally lower practical thresholds for approval, potentially normalising broader access beyond what lawmakers initially intended. Vulnerable groups could experience indirect pressure where technological efficiency reshapes clinical practice norms. Although statutory limits may remain formally unchanged, automation may produce functional expansion without corresponding legislative amendment, revealing a gap in oversight of technological impact on application.

Liability and accountability further illustrate regulatory fragmentation. If an AI system contributes to an erroneous assessment or malfunctions during implementation, responsibility becomes difficult to attribute23. Traditional malpractice and tort regimes presuppose identifiable human actors, yet AI-mediated processes involve developers, platform providers, device manufacturers, and clinicians. Current legal doctrines do not clearly allocate liability in such hybrid human-machine contexts, leaving uncertainty regarding accountability for technological errors.

Data governance also presents regulatory challenges. AI systems require extensive access to sensitive health information, increasing risks of data breaches, secondary data use24, and cross-border storage issues. While general data protection laws may apply, they rarely contain provisions tailored specifically to AI use in assisted dying contexts. The absence of targeted governance standards for data processing in this sensitive domain further demonstrates regulatory incompleteness.

Beyond procedural concerns lies the broader structural risk of commercialization and technological instrumentalization. As automated tools and self-executing devices emerge, assisted dying technologies may increasingly develop within market-driven frameworks. Current regulatory regimes seldom address procurement standards, certification requirements, or advertising controls for such technologies. Without oversight, commercial incentives could shape development priorities in ways that emphasise technological efficiency over patient protection. This gap highlights the need for regulatory scrutiny of economic dynamics alongside technical safeguards.

Taken together, technological integration exposes weaknesses in regulatory systems that were designed for human-centred decision-making. The expansion of AI-mediated evaluation, automated facilitation, and speculative technological systems reveals gaps in algorithmic governance, liability allocation, misinformation control, autonomy and informed consent, procedural oversight, data protection, and market regulation. Addressing these deficiencies requires proactive legal adaptation rather than reactive governance after technological deployment.

RECOMMENDATIONS

In light of the regulatory gaps exposed by technological integration in euthanasia and assisted suicide, legal frameworks must move toward anticipatory governance rather than reactive intervention. Legislatures should clarify the permissible scope of technological participation in end-of-life decision-making by explicitly defining whether and how artificial intelligence systems may be used in capacity assessments, eligibility evaluation, and procedural verification.

Regulatory authorities should also establish clear classification standards for AI tools deployed in this context, determining whether such systems fall under medical device regulation or require specialised oversight mechanisms. Transparency requirements, algorithmic audit obligations, and mandatory human supervision thresholds should be introduced to ensure that technological systems remain subordinate to professional accountability.

Furthermore, liability frameworks must be updated to address hybrid human-machine decision-making. Clear rules should delineate responsibility among clinicians, developers, manufacturers, and platform providers to prevent accountability fragmentation. Data governance standards specific to AI use in assisted dying should be developed to safeguard sensitive patient information and minimise privacy risks.

Finally, regulators should consider mechanisms to prevent commercialisation from undermining patient-centred care. Certification standards, procurement oversight, and restrictions on promotional practices may be necessary to ensure that technological innovation does not transform assisted dying into a market-driven enterprise detached from its regulatory safeguards.

CONCLUSION

Technological advancement has transformed assisted dying from a primarily physician-centred practice into one increasingly influenced by automated systems, digital tools, and speculative engineering concepts. While these innovations may appear to enhance efficiency or autonomy, they simultaneously expose structural limitations within existing regulatory frameworks. Without anticipatory legal adaptation, technological integration risks widening accountability gaps and weakening procedural safeguards. Accordingly, regulatory systems must evolve in parallel with technological development to ensure that human oversight, legal responsibility, and patient protection remain central to assisted dying governance.

REFERENCES

Kalaivani Annadurai, Raja Danasekaran and Geetha Mani, 'Euthanasia: Right to Die with Dignity' (2014) 3(4) Journal of Family Medicine and Primary Care 477-478 https://pmc.ncbi.nlm.nih.gov/articles/PMC4311376/ accessed 27 February 2026. Bernard M Dickens, Joseph M Boyle and Linda Ganzini, 'Euthanasia and Assisted Suicide' in The Cambridge Textbook of Bioethics (edited by Peter A Singer and A M Viens, Cambridge University Press 2008) 72-77 https://www.cambridge.org/core/books/cambridge-textbook-of-bioethics/euthanasia-and-assisted-suicide/7C7CCC8B2BF820A5AF8643FB3C0C6048 accessed 27 February 2026. Yelson Alejandro Picon-Jaimes and others, 'Euthanasia and assisted suicide: An in-depth review of relevant historical aspects' (2022) 75 Annals of Medicine and Surgery 103380, doi:10.1016/j.amsu.2022.103380 https://pmc.ncbi.nlm.nih.gov/articles/PMC8857436/ accessed 27 February 2026. Wikipedia, 'Legality of euthanasia' (Wikipedia) https://en.wikipedia.org/wiki/Legality_of_euthanasia accessed 27 February 2026. Miriam Colleran and Anne M Doherty, 'Examining Assisted Suicide and Euthanasia Through the Lens of Healthcare Quality' (2024) 193(1) Irish Journal of Medical Science 353-362, doi:10.1007/s11845-023-03418-2 https://pmc.ncbi.nlm.nih.gov/articles/PMC10808165/ accessed 27 February 2026. Ijaz Durrani, 'AI-Assisted Euthanasia: Ethical, Technological, and Social Considerations' (ResearchGate, November 2024) https://www.researchgate.net/publication/385619633_AI-Assisted_Euthanasia_Ethical_Technological_and_Social_Considerations accessed 3 March 2026. M Gulino and others, 'How Can Euthanasia and Assisted Suicide Regulation Guarantee Patient Health and Autonomy? Lessons from Nine European Countries' (2025) 33 Ethics, Medicine and Public Health 101131 https://www.sciencedirect.com/science/article/abs/pii/S2352552525000908 accessed 3 March 2026. J Pereira, 'Legalizing Euthanasia or Assisted Suicide: The Illusion of Safeguards and Controls' (2011) 18(2) Current Oncology e38-e45, doi:10.3747/co.v18i2.883 https://pmc.ncbi.nlm.nih.gov/articles/PMC3070710/ accessed 3 March 2026. Brian Martin, 'Techniques to Pass On: Technology and Euthanasia' (2010) 30(1) Bulletin of Science, Technology & Society 54-59, doi:10.1177/0270467609355053 https://www.researchgate.net/publication/249747622_Techniques_to_Pass_on_Technology_and_Euthanasia accessed 3 March 2026. Madison Frank, 'Dignity in Death: Why the Fundamental Right of the Individual to Choose Their Final Moments Outweighs the Government's Societal Interests to Preserve Life' (2020) 49(2) University of Baltimore Law Review https://scholarworks.law.ubalt.edu/ublr/vol49/iss2/2 accessed 3 March 2026. Exit International, 'Installations' (Exit International) https://www.exitinternational.net/about-exit/installations/ accessed 3 March 2026. Ibid Ibid NDTV News Desk, 'Controversial Sarco "Suicide" Pods New AI Feature to Allow Couples to Die Together' (NDTV, 20 January 2026) https://www.ndtv.com/feature/controversial-sarco-suicide-pods-new-ai-feature-to-allow-couples-to-die-together-10794861 accessed 3 March 2026. Wikipedia, 'Euthanasia Coaster' (Wikipedia) https://en.wikipedia.org/wiki/Euthanasia_Coaster accessed 3 March 2026. The Museum of Modern Art, 'Euthanasia Coaster -- Julijonas Urbonas' (MoMA) https://www.moma.org/interactives/exhibitions/2013/designandviolence/euthanasia-coaster-julijonas-urbonas/ accessed 3 March 2026. CARE UK, 'Assisted suicide implant invented for dementia patients' (CARE UK, 4 June 2025) https://care.org.uk/news/2025/06/new-kill-switch-implant-proposed-for-early-stage-dementia-patients accessed 3 March 2026. Paulo Bolinhas, 'Artificial Intelligence and health disinformation: A case study on euthanasia and the new communication processes' (ResearchGate, December 2024) https://www.researchgate.net/publication/395732596_Artificial_Intelligence_and_health_disinformation_A_case_study_on_euthanasia_and_the_new_communication_processes accessed 4 March 2026. Ibid Abiodun Adegbesan and others, 'Ethical Challenges in the Integration of Artificial Intelligence in Palliative Care' (2024) 4 Journal of Medicine, Surgery, and Public Health 100158, doi:10.1016/j.glmedi.2024.100158 https://www.sciencedirect.com/science/article/pii/S2949916X24001117 accessed 4 March 2026. Durrani (n6) Ibid Hannah van Kolfschooten, 'The prospects of using AI in euthanasia and physician-assisted suicide: a legal exploration' (2024) 5 AI and Ethics 1461-1466, doi:10.1007/s43681-024-00491-w https://link.springer.com/article/10.1007/s43681-024-00491-w accessed 4 March 2026. Bolinhas (n18)