Introduction: The Nexus of AI and Forensic Science

The field of forensic science is undergoing a fundamental transformation, shifting from traditional, labor-intensive analog methods to sophisticated digital systems augmented by artificial intelligence (Montgomery, 2025a). This technological evolution is driven by AI's capacity to process vast datasets with speed and precision, learn intricate patterns, and automate tasks that once required significant human effort (Ahmed Alaa El-Din, 2022). The application of AI has made its entrance into virtually every aspect of a criminal investigation, from the initial analysis of a crime scene to the presentation of evidence in court (Cărcăle, 2025a).

The promise of AI in this context is compelling. AI-driven tools can drastically reduce processing times for essential forensic tasks and enhance the accuracy of evidence analysis by minimizing the potential for human subjectivity and error (Enhancing the Evidence with Algorithms, 2025). For example, AI can analyze complex DNA mixtures, reconstruct crime scenes using advanced modeling, and assist in ballistics and fingerprint analysis with unprecedented efficiency (Cărcăle, 2025b).

The rapid integration of these technologies into the justice system has outpaced the development of legal and ethical frameworks to govern their use. The central question is a profound one: how can legal systems, which are built on principles of transparency, due process, and the right to challenge evidence, effectively and justly integrate technologies that are often opaque, prone to bias, and challenging to validate?

This report provides a comprehensive analysis of the legal and ethical landscape governing AI in forensic evidence, focusing on three critical areas: the foundational legal standards for admissibility, the key challenges posed by AI's unique nature, and a comparative analysis of international legal frameworks (Montgomery, 2025a). The ultimate aim is to demonstrate that a new, interdisciplinary legal and ethical paradigm is urgently needed to align technological progress with the fundamental principles of justice and human rights (Ahmed Alaa El-Din, 2022).

Part I: The Role of AI in Modern Forensic Practice

A. Current Applications and Transformative Potential

AI is no longer a theoretical concept in forensic science; it is actively being deployed to support, augment, and even replace traditional methods in a variety of disciplines (Cărcăle, 2025c).

• Forensic Medicine and Biometrics: AI enhances forensic identification by analyzing vast databases in seconds, a task that would take human experts months or even years. AI-driven tools for facial recognition, fingerprint matching, and biometric analysis improve the precision and speed of identifying victims and suspects ( AI in Forensic Science, 2025) . Beyond identification, AI assists in the assessment of traumatic injuries and the estimation of the post-mortem interval, helping to reconstruct crime timelines and assess the nature and cause of wounds. (Digital Forensics - Sensity AI, 2025).
• Digital Forensics: The proliferation of digital data has made digital forensics a cornerstone of modern investigations. AI-powered tools are essential for analyzing huge datasets from emails, social media, and the dark web to detect fraud or cyber threats.These systems can decrypt files and recover deleted data using deep learning algorithms (Montgomery, 2025b). A critical application is the detection of deepfakes and other manipulated digital content, which traditional tools often fail to identify. AI can authenticate digital evidence, trace its origin, and verify its integrity before it is presented in court (The Growing Role of AI in Criminal Justice, 2025).
• Pattern Recognition & Crime Scene Analysis: AI excels at recognizing patterns that may be invisible or unmanageable for a human analyst. In ballistics, AI systems can match bullets and shell casings to specific weapons with high efficiency (Novelli et al., 2024). For crime scene reconstruction, AI can use various inputs to create animated videos that suggest different scenarios, greatly assisting human experts. Furthermore, AI can improve the quality of low-resolution images and videos to identify individuals or objects, leveraging advanced mathematical representations to make identifications that were previously impossible (The Growing Role of AI in Criminal Justice, 2025).
• DNA & Molecular Biology: AI accelerates DNA analysis by identifying genetic markers in minutes and can be used for predictive phenotyping, which involves estimating a suspect's appearance from DNA (AI in Forensic Science, 2025; BCIT & Chen, 2025). More critically, AI algorithms can decipher complex DNA mixtures containing multiple donors, a task that is often too chaotic for manual analysis. AI-based data mining separates and identifies individual DNA profiles from these large, complex datasets, making previously unusable DNA evidence viable for legal proceedings (BCIT & Chen, 2025).

B. A Tool, not a Replacement: The Hybrid Approach

A unifying principle across the current landscape of AI in forensics is that it functions optimally as a supplementary tool, rather than a total replacement for human expertise. This symbiotic relationship is increasingly being adopted as the gold standard in the field (Farber, 2025a).

AI systems are excellent at performing the initial, rapid screening of large volumes of data, acting as advanced filters to narrow down possible matches or patterns for human review. The human forensic expert, meanwhile, brings irreplaceable insights, qualitative judgment, and contextual understanding that an algorithm cannot replicate (Farber, 2025b). A machine can identify a pattern, but a human expert must interpret what that pattern means within the context of the specific crime and legal framework. This is a critical legal and ethical consideration. The fact that an expert witness was unable to explain how an AI-enhanced image was generated, leading to the evidence being thrown out of court, illustrates the judicial system's reliance on the human-in-the-loop model (BCIT & Chen, 2025). The courts require a human expert who can be cross-examined and who can explain the AI's process and its limitations, rather than accepting a conclusion from an opaque, automated system. The final analysis and ultimate decision-making must remain with the human expert and the trier of fact (Montgomery, 2025c).

Part II: Admissibility of Foundational Legal Standards for Scientific Evidence

The admissibility of scientific evidence in a courtroom is not a matter of a judge's personal discretion; it is governed by established legal standards that have evolved over the last century (Montgomery, 2025c). AI-generated evidence, as a form of expert testimony, must adhere to these same principles.

A. The Frye Standard: The General Acceptance Test

The Frye standard, established in the 1923 case of Frye v. United States, focuses on a single criterion for admitting expert opinion evidence: "general acceptance" within the relevant scientific community (BurnsWhite, 2025). This test emerged from a case involving a primitive lie detector, where the court ruled the technology was not yet sufficiently established to be reliable (Kelly-Frye, Daubert, 2025).

Under Frye, the court's role is relatively passive. The judge's inquiry is not into the reliability of the scientific principle itself, but whether the methodology from which the expert's opinion is deduced has been "sufficiently established to have gained general acceptance". This standard is still followed by some state courts today. A notable example is People v. Wakefield, in which a court applying the Frye standard found the TrueAllele probabilistic genotyping software admissible (BurnsWhite, 2025). The court determined that the software, which uses a degree of AI, was "generally accepted" based on a "plethora of evidence" and a lack of significant evidence to the contrary.

B. The Daubert Standard: The "Gatekeeping" Role of the Judge

In 1993, the U.S. Supreme Court's decision in Daubert v. Merrell Dow Pharmaceuticals, Inc. superseded the Frye standard for federal courts and was subsequently adopted by many states.The Daubert ruling transformed the judge from a passive observer of scientific consensus into an active gatekeeper, charged with ensuring that expert testimony rests on a reliable foundation and is relevant to the case (Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), 2025).

The Daubert standard, while more flexible than Frye, presents a significant challenge for AI-based forensic evidence. The requirement for a known or potential error rate and the existence of standards for operation directly conflict with the "black box" nature of many AI systems (Burns & White, 2025). The adversarial system allows attorneys to attack the lack of peer review, the secrecy of proprietary algorithms, or the testability of a model's internal workings.This friction between a legal framework built on principles of transparency and an inherently opaque technology is a core tension in the modern legal landscape (Evaluating the Use of AI in Digital Evidence and Courtroom Admissibility, Magnet Forensics, 2025).

Table 1: Comparative Analysis of Frye vs. Daubert Standards

Part III: Key Legal and Ethical Challenges of AI-Assisted Evidence

The integration of AI into the justice system introduces a new set of complex challenges that go beyond the foundational standards for expert testimony. These issues strike at the core of due process and equal protection under the law (Montgomery, 2025c).

A. The Black Box Problem

The black box problem is a term for the opaque nature of many AI systems, particularly deep learning models, whose decision-making processes are too complex for human comprehension. The internal workings of these algorithms are enigmatic and opaque (University of Michigan-Dearborn, 2025).

This problem is not a simple matter of forgetting; rather, the system was never designed to track which specific inputs led to a conclusion.This lack of explainability is a direct impediment to legal proceedings. The legal implications of this problem are profound. A defendant's fundamental right to confront and challenge the evidence against them is undermined when the technology's conclusions cannot be explained or independently verified (Black Box Technologies Used in Criminal Prosecutions, 2025). Furthermore, courts often trust these proprietary technologies and may even protect them from closer examination. In one case, a judge denied a defense request for independent evaluators to review DNA technology because transparency would prevent the company from marketing its product (Black Box Technologies Used in Criminal Prosecutions, 2025). This action reveals a profound conflict: the legal system's reliance on private companies' secretive and complex science can prioritize corporate interests and intellectual property over a defendant's constitutional rights. The lack of transparency has led to the exclusion of evidence in court, as seen in a case where an expert witness could not explain how an AI tool had generated an enhanced image, resulting in the evidence being thrown out entirely.

B. Algorithmic and Data Bias

A central ethical paradox of AI in forensics is its potential to reduce human subjectivity and cognitive bias while simultaneously introducing its own systemic and structural biases. These algorithmic biases can be particularly dangerous because they are hidden under a veil of objectivity, perpetuating or even exacerbating existing racial, socioeconomic, and gender disparities (Jinad et al., 2024).

The use of AI risk assessment tools in bail and sentencing decisions illustrates this dynamic perfectly. While these tools are presented as objective, they often use proxies such as zip codes or employment status that correlate with race and class, thereby penalizing marginalized groups and institutionalizing discrimination (Jinad et al., 2024). The tension between the promise of "algorithmic justice" and the reality of "digital discrimination" is one of the most pressing ethical concerns in the field. The question becomes: can we accept a system where a few mistakes are deemed acceptable for the sake of efficiency, when those mistakes result in the sacrifice of an individual's life and liberty (Jinad et al., 2024).

Table 2.2: Types of Bias in AI-Assisted Forensics

C. Authentication and Admissibility of AI-Generated Evidence

The proliferation of generative AI, which can create highly realistic deepfakes of videos and audio, has brought about a crisis of confidence in the authenticity of digital media. This presents a new legal challenge where defence attorneys may increasingly assert AI fakery as a form of reasonable doubt, even without technical proof (Artificial Intelligence and the Law, 2025; Metallo, 2025).

This phenomenon, dubbed the AI doubt effect, can cause jurors to overestimate the plausibility of fabrication, eroding trust in even genuine evidence, much like the CSI effect. To combat this, courts are assuming a gatekeeping function to prevent speculative doubt from supplanting procedural rigor. Legal frameworks are adapting to these new threats. For instance, courts are beginning to require proponents of digital evidence to demonstrate a clear chain of custody and to provide expert validation (Metallo, 2025).

The Federal Rules of Evidence, particularly Rule 702 and Rule 901, are being applied to AI-assisted evidence. Rule 702 requires that an expert's testimony be based on sufficient facts and data, and that it is the product of reliable principles and methods. Rule 901(b)(9) provides a mechanism for authenticating evidence by describing a process or system and showing that it produces an accurate result (Artificial Intelligence and the Law, 2025). To fulfill this, a witness with personal knowledge of how the AI tool functions must be able to describe the process and prove its accuracy, which makes Explainable AI a legal necessity. Proposed rule changes, such as a 2023 amendment to Rule 702, have clarified that judges must apply a preponderance of the evidence standard when determining whether these requirements have been met (Artificial Intelligence and the Law, 2025).

Part IV: Comparative Legal Approaches to AI in Forensic Evidence

Jurisdictions around the world are grappling with the legal and ethical implications of AI with varying approaches. While the United States relies on its existing common law and a fragmented, reactive approach, other regions are adopting proactive regulatory frameworks (Artificial Intelligence and the Law, 2025).

A. United States: A Jurisdictional Patchwork

The U.S. currently lacks a unified, federal regulatory framework to govern AI. Instead, its governance is a patchwork of federal and state-level initiatives, as well as voluntary industry standards. The legal system relies on the continued, and sometimes inconsistent, application of the Frye and Daubert standards on a case-by-case basis as AI-related evidence is challenged in court (BurnsWhite, 2025). This reactive, litigation-driven approach means that legal precedents are only established after a technology has been deployed and its potential for harm has been litigated (Kelly-Frye, Daubert, Mohan, 2025).

B. European Union: The EU AI Act and a Risk-Based Framework

In contrast to the U.S. approach, the European Union has adopted the EU AI Act, the world's first comprehensive legal framework on AI. The Act takes a unified, human-centric, and proactive approach by categorizing AI systems based on a risk-based model, from unacceptable to high-risk, limited-risk, and minimal-risk (Key Insights into AI Regulations in the EU and the US, 2025).

AI systems used in law enforcement and the administration of justice are classified as high-risk because they can interfere with an individual's fundamental rights (AI Act, Shaping Europe’s Digital Future, 2025). These high-risk systems are subject to stringent obligations, which developers and deployers must meet before the systems are even placed on the market. The EU's proactive framework is designed to prevent harm before it occurs, a notable departure from the U.S. litigation model.

The EU is building a legal foundation to foster trustworthy AI from the outset, although there are still concerns about the permissiveness of the Act in areas like counter-terrorism (admin gnet, 2025).

C. The United Kingdom and Canada: Evolving Common Law Principles

Both the UK and Canada, with their common law traditions, have legal frameworks for expert testimony that have been influenced by U.S. jurisprudence but retain their own unique characteristics (Expert Evidence, The Crown Prosecution Service, 2025).

In the United Kingdom, expert evidence is admissible if it is necessary to assist the court, comes from a qualified and impartial expert, and is based on a sufficiently reliable scientific basis (Booth et al., 2019). The Privy Council's guidance on expert evidence, established in Lundy v R, includes factors for reliability that are almost identical to the Daubert factors: testability, peer review, known error rate, and general acceptance (Legal Obligations, 2025).In Canada, the framework for expert evidence is defined by the Supreme Court (SC) case R. v. Mohan. This standard requires expert testimony to be relevant, necessary, not subject to an exclusionary rule, and from a properly qualified expert (Glancy & Bradford, 2007). The standard of necessity is a high bar, requiring that the evidence be beyond the experience and knowledge of a judge or jury.For a scientific theory or technique, the evidence is subject to a special scrutiny to ensure it meets a basic reliability threshold, a clear echo of the Daubert standard (Glancy & Bradford, 2007).

D. The German Inquisitorial System

The German legal system offers a compelling contrast to the adversarial models in the U.S., UK, and Canada. In Germany, the system is inquisitorial, meaning the court, rather than the parties' lawyers, holds the primary responsibility for gathering and sifting evidence (Booth et al., 2019).

This procedural difference fundamentally changes the dynamic surrounding the use of AI. The court appoints and directs its own experts to prepare reports. Experts retained by the parties do not carry the same legal weight as formal evidence (Schmeilzl, 2018). This model could offer a procedural solution to the black box problem and the conflict with corporate trade secrets. Instead of a defense lawyer having to compel a private company to disclose its proprietary secrets, the court itself has the authority to direct its own expert to ensure the reliability of the evidence it relies on (Taking of Evidence, European E-Justice Portal, 2025). This approach places the burden of ensuring evidence reliability on the court as a matter of judicial responsibility, rather than on the parties as a matter of litigation strategy (Patent System of Germany - Judicial Patent Proceedings, 2025).

Table 2.3: Comparative Analysis of International Legal Frameworks

Part V: Future Outlook and Recommendations

A. The Path to Explainable AI (XAI)

The long-term solution to the black box problem is a collaborative effort between technology and law. The development of Explainable AI (XAI) is critical, as it focuses on creating AI models that can deliver accurate, meaningful, and understandable explanations for their outputs. This is not merely a technical challenge; it is a legal and ethical imperative (Ciligot, 2023). Systems must be built with built-in accountability, logging the AI's reasoning process, its confidence levels, and the datasets used to train it, to allow for independent verification and legal scrutiny (Ciligot, 2023).

B. Legal and Policy Recommendations

Current legal standards, which were designed for a different era of science, are insufficient for the unique challenges of AI. New legal frameworks are needed to address issues of algorithmic bias, transparency, and data privacy (Veiligheid, 2024). Jurisdictions should consider:

New Rules of Evidence:

The development of specific rules for AI evidence, such as a proposed Rule 707, could provide clear guidance on admissibility standards and disclosure requirements(Veiligheid, 2024).

Interdisciplinary Collaboration:

To create effective and culturally responsive guidelines, policymakers, legal scholars, technologists, and forensic experts must work together.

Mandatory Audits and Impact Assessments:

 To proactively identify and mitigate biases before they can cause harm, high-risk AI systems used in criminal justice should be subject to mandatory algorithmic impact assessments and independent audits (Artificial Intelligence and the Law, 2025).

C. The Human Element in a Data-Driven World

The central theme of this report is that AI is a tool, not a replacement for human expertise. While AI can enhance efficiency and accuracy, the final interpretation and judgment of evidence must remain with the human expert and the trier of fact (Cărcăle, 2025d). Justice cannot be outsourced to a machine.The integrity of the justice system relies on an unwavering commitment to equity and the human capacity for reasoned, qualitative judgment.

Conclusion

The integration of AI into forensic science holds immense potential to enhance the pursuit of justice, but its full realization is complicated by foundational legal standards, the "black box" problem, and insidious algorithmic biases. Legal systems worldwide are grappling with these challenges with varying degrees of success, with the EU's proactive regulatory approach standing in stark contrast to the U.S.'s reactive, litigation-driven model. The path forward requires a fusion of rigorous legal oversight, a commitment to developing transparent and fair AI, and a renewed appreciation for the irreplaceable role of human expertise and ethical judgment. The ultimate success of AI in justice will depend not on the code itself, but on the values, we encode and the safeguards we put in place to protect fundamental human rights. As courts refine their focus to discern the real from the fabricated in a world of ever-advancing AI, they must apply historical wisdom and procedural rigor to ensure that technology serves justice, rather than undermining it.

Reference:

• admin_gnet. (2025, March 10). The EU’s AI Act: Implications on Justice and Counter-Terrorism. GNET. https://gnet-research.org/2025/03/10/the-eus-ai-act-implications-on-justice-and-counter-terrorism/
• Ahmed Alaa El-Din, E. (2022). ARTIFICIAL INTELLIGENCE IN FORENSIC SCIENCE: INVASION OR REVOLUTION? Egyptian Society of Clinical Toxicology Journal, 10(2), 20–32. https://doi.org/10.21608/esctj.2022.158178.1012
• AI Act | Shaping Europe’s digital future. (2025, August 1). https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
• AI in Forensic Science: How Machine Learning is Revolutionizing Crime Investigation – Do Forensics. (n.d.). Retrieved August 25, 2025, from https://doforensics.se/ai-in-forensic-science-how-machine-learning-is-revolutionizing-crime-investigation/
• AI in Forensic Science: How Machine Learning is Revolutionizing Crime Investigation – Do Forensics. (n.d.). Retrieved August 25, 2025, from https://doforensics.se/ai-in-forensic-science-how-machine-learning-is-revolutionizing-crime-investigation/
• AI’s mysterious ‘black box’ problem, explained | University of Michigan-Dearborn. (n.d.). Retrieved August 25, 2025, from https://umdearborn.edu/news/ais-mysterious-black-box-problem-explained
• Artificial Intelligence and the Law: Expert Witness Issues in 2025. (n.d.-a). Retrieved August 25, 2025, from https://www.forensisgroup.com/resources/expert-legal-witness-blog/artificial-intelligence-law
• Artificial Intelligence and the Law: Expert Witness Issues in 2025. (n.d.-b). Retrieved August 25, 2025, from https://www.forensisgroup.com/resources/expert-legal-witness-blog/artificial-intelligence-law
• Artificial Intelligence and the Law: Expert Witness Issues in 2025. (n.d.-c). Retrieved August 25, 2025, from https://www.forensisgroup.com/resources/expert-legal-witness-blog/artificial-intelligence-law
• BCIT, & Chen, A. (2025, July 24). The future of AI in forensic science—BCIT. https://www.bcit.ca/news/stories/the-future-of-ai-in-forensic-science/
• Black Box Technologies Used in Criminal Prosecutions. (n.d.). Retrieved August 25, 2025, from https://www.salazarandkelly.com/black-box-technologies-used-in-criminal-prosecutions/
• Booth, B. D., Watts, J., & Dufour, M. (2019). Lessons from Canadian Courts for All Expert Witnesses. Journal of the American Academy of Psychiatry and the Law Online. https://doi.org/10.29158/JAAPL.003838-19
• BurnsWhite. (2025, March 3). The Differences Between the Frye & the Daubert Standards. Burns White. https://burnswhite.com/differences-between-frye-daubert/
• Cărcăle, V.-A. (2025a). The Future of Artificial Intelligence (AI) Applications in Forensics. https://doi.org/10.5281/ZENODO.15481632
• Cărcăle, V.-A. (2025b). The Future of Artificial Intelligence (AI) Applications in Forensics. https://doi.org/10.5281/ZENODO.15481632
• Cărcăle, V.-A. (2025c). The Future of Artificial Intelligence (AI) Applications in Forensics. https://doi.org/10.5281/ZENODO.15481632
• Cărcăle, V.-A. (2025d). The Future of Artificial Intelligence (AI) Applications in Forensics. https://doi.org/10.5281/ZENODO.15481632
• Ciligot, C. (2023, October 30). Evaluating the use of AI in digital evidence and courtroom admissibility. Magnet Forensics. https://www.magnetforensics.com/blog/evaluating-the-use-of-ai-in-digital-evidence-and-courtroom-admissibility/
• Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). (n.d.). Justia Law. Retrieved August 25, 2025, from https://supreme.justia.com/cases/federal/us/509/579/
• Digital Forensics—Sensity AI. (n.d.). Retrieved August 25, 2025, from https://sensity.ai/use-cases/digital-forensics/
• Enhancing the Evidence with Algorithms: How Artificial Intelligence Is Transforming Forensic Medicine. (n.d.). Retrieved August 25, 2025, from https://www.mdpi.com/2075-4418/13/18/2992
• Evaluating the use of AI in digital evidence and courtroom admissibility—Magnet Forensics. (n.d.). Retrieved August 25, 2025, from https://www.magnetforensics.com/blog/evaluating-the-use-of-ai-in-digital-evidence-and-courtroom-admissibility/
• Expert Evidence | The Crown Prosecution Service. (n.d.). Retrieved August 25, 2025, from https://www.cps.gov.uk/legal-guidance/expert-evidence
• Farber, S. (2025a). AI as a decision support tool in forensic image analysis: A pilot study on integrating large language models into crime scene investigation workflows. Journal of Forensic Sciences, 70(3), 932–943. https://doi.org/10.1111/1556-4029.70035
• Farber, S. (2025b). AI as a decision support tool in forensic image analysis: A pilot study on integrating large language models into crime scene investigation workflows. Journal of Forensic Sciences, 70(3), 932–943. https://doi.org/10.1111/1556-4029.70035
• Glancy, G. D., & Bradford, J. M. W. (2007). The Admissibility of Expert Evidence in Canada. Journal of the American Academy of Psychiatry and the Law Online, 35(3), 350–356.
• Jinad, R., Gupta, K., Simsek, E., & Zhou, B. (2024). Bias and fairness in software and automation tools in digital forensics. Journal of Surveillance, Security and Safety, 5(1), 19–35. https://doi.org/10.20517/jsss.2023.41
• Kelly-Frye, Daubert, Mohan, and Why You Need to Understand These Cases. (n.d.). Axon.Com. Retrieved August 25, 2025, from https://www.axon.com/resources/kelly-frye-daubert-mohan-and-why-you-need-to-understand-these-cases
• Key insights into AI regulations in the EU and the US: Navigating the evolving landscape. (n.d.). Kennedys Law. Retrieved August 25, 2025, from https://kennedyslaw.com
• Legal obligations: Issue 8 (accessible). (n.d.). GOV.UK. Retrieved August 25, 2025, from https://www.gov.uk/government/publications/legal-obligations-issue-8/legal-obligations-issue-8-accessible
• Metallo, V. N. A. (n.d.). The Impact of Artificial Intelligence on Forensic Accounting and Testimony—Congress Should Amend “The Daubert Rule” to Include a New Standard.
• Montgomery, R. M. (2025a). Augmenting Forensic Science Through AI: The Next Leap in Multidisciplinary Approaches (No. 2025011951). Preprints. https://doi.org/10.20944/preprints202501.1951.v1
• Montgomery, R. M. (2025b). Augmenting Forensic Science Through AI: The Next Leap in Multidisciplinary Approaches (No. 2025011951). Preprints. https://doi.org/10.20944/preprints202501.1951.v1
• Montgomery, R. M. (2025c). Augmenting Forensic Science Through AI: The Next Leap in Multidisciplinary Approaches (No. 2025011951). Preprints. https://doi.org/10.20944/preprints202501.1951.v1
• Novelli, C., Casolari, F., Rotolo, A., Taddeo, M., & Floridi, L. (2024). AI Risk Assessment: A Scenario-Based, Proportional Methodology for the AI Act. Digital Society, 3(1), 13. https://doi.org/10.1007/s44206-024-00095-1
• Patent System of Germany—Judicial Patent Proceedings. (n.d.). Retrieved August 25, 2025, from https://www.wipo.int/patent-judicial-guide/en/full-guide/germany/5.6.7
• Schmeilzl, B. (2018, October 22). Testimony and Evidence in German Civil Litigation. GermanCivilProcedure.Com. https://www.germancivilprocedure.com/testimony-and-evidence-in-german-civil-litigation/
• Taking of evidence | European e-Justice Portal. (n.d.). Retrieved August 25, 2025, from https://e-justice.europa.eu/topics/court-procedures/civil-cases/taking-evidence/de_en
• The Growing Role of AI in Criminal Justice. (n.d.). MastersInAI.Org. Retrieved August 25, 2025, from https://www.mastersinai.org/industries/criminal-justice/
• Veiligheid, M. van J. en. (2024, February 26). Worldwide standard for forensic analysis laboratories—News item—Netherlands Forensic Institute [Nieuwsbericht]. Ministerie van Justitie en Veiligheid. https://www.forensicinstitute.nl/news/news/2024/02/26/worldwide-standard-for-forensic-analysis-laboratories