AI Liability and Cross-Border Sovereignty: Case Commentaries

by

Andrei-Ioan Muresan 

CyJurII Scholar

on 23 May 2026

Introduction

As artificial intelligence technologies seamlessly integrate into everyday life, global legal frameworks are facing unprecedented pressure to adapt. The intersection of autonomous systems and traditional legal structures raises profound questions about accountability, corporate negligence, intellectual property ownership, and the ethical responsibilities of professional practitioners.

This commentary analyzes three landmark international legal cases that highlight how modern jurisprudence interacts with AI:

• Product & Corporate Liability: Evaluating how autonomous driving failures shift accountability from the operator to the manufacturer under established product liability doctrines.

• Intellectual Property Rights: Assessing whether AI systems can possess legal personality and be recognized as autonomous inventors under existing patent frameworks.

• Professional and Ethical Responsibility: Addressing the legal risks of "AI hallucinations" in corporate litigation and reinforcing a practitioner’s strict "Duty to Know."

Together, these cases illustrate the shifting boundary lines of liability and sovereignty in an AI-driven global landscape.

Case Commentaries

1. Benavides v. Tesla Inc.

Court: U.S. District Court, Southern District of Florida (No. 21-21940)

Case Background

The legal saga originated from a 2019 vehicular crash in Key Largo, Florida. George McGee, the owner of a Tesla Model S, collided at 65 miles per hour into a parked Chevy Tahoe while actively operating the vehicle on Tesla's "Autopilot" system. The high-speed impact resulted in the tragic death of 22-year-old Naibel Benavides and caused severe, debilitating injuries to her boyfriend, Dillon Angulo. In response, Benavides’ family alongside Angulo initiated legal proceedings against Tesla Inc..

Legal Arguments & Claims

The plaintiffs sought to hold Tesla accountable for the crash by applying strict product liability concepts. Their actions targeted the manufacturer under four primary legal headings:

• Defective design

• Failure to warn

• Defective manufacture

• Negligent misrepresentation

Judicial Rulings & Final Judgment

In June 2025, the court issued an initial ruling dismissing the claims of defective manufacture and negligent misrepresentation. However, the court permitted the defective design and failure to warn claims to advance to trial.

In its final judgment, the court allocated the majority of the fault to the driver, George McGee, concluding that he was distracted and looking for his mobile phone immediately prior to the crash while relying on the Autopilot system. Crucially, however, the court also found Tesla legally at fault under both the defective design and failure to warn doctrines.

Legal Significance

Benavides v. Tesla Inc. serves as a monumental milestone because it directly pits the emerging concepts of AI liability against time-tested, traditional tort doctrines. The judgment highlighted two fatal flaws in Tesla's system:

1. Failure to Warn: Tesla failed to adequately warn vehicle owners within their operational manuals regarding the inherent risks of deploying "Autopilot".

2. Defective Design: The Autopilot system was poorly designed, allowing drivers to engage autonomous features on incompatible roadways and failing to provide critical alert notifications to drivers prior to an imminent collision.

2. Thaler v. Vidal

Court: U.S. Court of Appeals for the Federal Circuit (43 F.4th 1207, 1210, Fed. Cir. 2022)

Case Background

This landmark cyber-law dispute centers around Dr. Stephen Thaler and his proprietary AI software system, named DABUS (Device for the Autonomous Bootstrapping of Unified Sentience). Dr. Thaler asserted that his AI software autonomously generated two distinct inventions: a flickering light designed to mimic neural activity, and a fractal-shaped beverage container engineered to assist robots in gripping it efficiently.

Legal Arguments & Claims

Dr. Thaler filed two patent ownership applications with the United States Patent and Trademark Office (USPTO). He attributed the sole credit for inventing these items to DABUS, arguing that despite being the owner of the software, the machine itself was the true creator.

The USPTO rejected both applications under the statutory guidelines of the Patent Act, prompting Dr. Thaler to sue the USPTO and its Director, Katherine K. Vidal, in the U.S. District Court for the Eastern District of Virginia. Following a ruling upholding the USPTO's stance, Thaler appealed the matter to the U.S. Court of Appeals for the Federal Circuit.

In his appellate plea, Thaler presented four arguments:

• The definition of an "inventor" under the Patent Act should be legally extended to encompass non-human entities.

• Patent protections should cover AI-generated inventions to foster "innovation and public disclosure".

• Recognizing AI systems as inventors directly supports the constitutional objective of promoting the progress of science and the useful arts.

• South Africa served as a global precedent, having already granted a patent recognizing his AI system as an inventor.

Judicial Rulings & Final Judgment

The Federal Circuit rejected the appeal, affirming the consistent positions held by both the USPTO and the U.S. District Court. The appellate court ruled that under prevailing statutory frameworks, inventors must be human beings.

Legal Significance

This case firmly established the boundary between AI liability, creation, and property rights. The legal precedent dictates two critical rules:

• Human Exclusivity: Inventions generated entirely by artificial intelligence machines cannot be patented as such under current U.S. patent law, which mandates that inventions must originate from an individual or a collective group of human individuals.

• Lack of Legal Personality: The court definitively affirmed that AI machines hold absolutely no legal personality or independent standing under the law.

3. Northbound Processing (Pty) Ltd v. South African Diamond and Precious Metals Regulator and Others

Court: High Court of South Africa, Gauteng Local Division, Johannesburg ([2025] ZAGPJHC 538)

Case Background

In January 2025, a commercial entity named Northbound Processing filed an urgent legal application seeking to compel the South African Diamond and Precious Metals Regulator to issue a refining license. The Regulator had withheld the license due to an ongoing internal shareholder dispute involving Rappa Resources—the entity that held the prior license. Northbound Processing countered that they had successfully satisfied all regulatory conditions for the license's release, including finalizing the structural business transfer and returning Rappa’s original license.

Legal Arguments & Claims

Because the Regulator steadfastly refused to issue the license, the standoff escalated to the High Court. Initially, the presiding judge evaluated the merits of the case and was prepared to grant the refining license to Northbound Processing.

However, during the formal drafting of the judgment, the judge uncovered a severe ethical violation: Northbound Processing's legal counsel had built their arguments upon fabricated case laws and fictional citations.

Judicial Rulings & Final Judgment

An investigation revealed that the fictional precedents were the product of AI hallucinations generated by a legal AI tool called "Legal Genius". Though the software was marketed as being trained exclusively on valid South African legal judgments and regional legislation, it invented non-existent case laws. As a direct consequence of presenting false data to the bench, the judge referred the offending lawyers to the Legal Practice Council for disciplinary proceedings.

Legal Significance

This decision is a cornerstone international precedent regarding the integration of AI tools within the legal profession. It directly juxtaposes AI capability against human professional accountability and the legal doctrine of the "Duty to Know".

• Permissibility: The ruling establishes that lawyers are entirely permitted to leverage generative AI utilities to streamline research and reinforce their arguments before a court of law.

• Verification Mandate: Crucially, the case underscores that artificial intelligence does not absolve human practitioners of liability. Legal professionals bear an uncompromisable, strict duty to meticulously analyze, verify, and vet every piece of AI-generated evidence to ensure its absolute legitimacy and validity before presenting it to a court.