Emerging Criminal Liability For Autonomous Vehicles And Transport Systems
1. Introduction – Criminal Liability and Autonomous Vehicles
As autonomous vehicles (AVs) increasingly operate with minimal or no human control, legal systems worldwide face the question:
Who is criminally liable when an autonomous vehicle causes harm?
Traditionally, criminal liability requires:
Actus reus (a guilty act) – a voluntary human act that causes harm, and
Mens rea (a guilty mind) – intention, recklessness, or negligence.
However, AVs complicate this:
The “driver” may not be controlling the vehicle.
The manufacturer or software designer may have indirect control.
The vehicle’s AI may make autonomous decisions unforeseen by its programmers.
This creates challenges for assigning culpability among:
Human users or “drivers”
Manufacturers and software developers
Maintenance companies
Vehicle owners
Even regulatory agencies (in rare cases)
2. Categories of Criminal Liability in AV Context
Driver/User Liability – when the person in the car fails to intervene or misuses the technology.
Manufacturer Liability – for defective design, inadequate testing, or failure to warn.
Software Developer Liability – for negligent coding, algorithmic bias, or malfunction leading to harm.
Corporate Criminal Liability – when corporate policies or cultures lead to systemic negligence.
Shared or Hybrid Liability – when multiple actors share responsibility.
3. Key Legal Theories Emerging
Negligence and Gross Negligence – failure to ensure proper safety or foresee risks.
Recklessness – deploying technology known to have unresolved safety issues.
Corporate Manslaughter – under statutes like the UK Corporate Manslaughter and Corporate Homicide Act 2007.
Product Liability – where the criminal component arises from knowing concealment or deliberate disregard for safety defects.
4. Important Case Law (Detailed Discussion of 6 Key Cases)
Case 1: State of Arizona v. Rafaela Vasquez (Uber Self-Driving Car Incident, 2018)
Jurisdiction: United States (Arizona)
Facts:
An Uber test vehicle (a Volvo XC90 modified for autonomous driving) struck and killed pedestrian Elaine Herzberg in Tempe, Arizona. The car was in self-driving mode, and the backup safety driver, Rafaela Vasquez, was watching videos on her phone instead of monitoring the road.
Legal Issue:
Was Vasquez criminally liable for negligent homicide even though the car was driving autonomously?
Decision & Reasoning:
Vasquez was charged with negligent homicide.
Prosecutors found that the software had recognized the pedestrian but failed to act appropriately due to algorithmic misclassification.
However, the human safety operator was found to have a duty to intervene.
Uber itself avoided criminal charges after cooperation and evidence that it had not intentionally neglected safety laws.
Significance:
This case demonstrates individual criminal liability even in semi-autonomous systems. It sets a precedent that human overseers of AVs remain responsible for active monitoring until systems achieve full autonomy.
Case 2: Tesla Autopilot Incidents – Various Cases (2016–2023)
Jurisdiction: United States, multiple states
Facts:
Multiple accidents have occurred while Tesla vehicles were operating in “Autopilot” or “Full Self-Driving (FSD)” mode, including fatal crashes. In some cases, drivers claimed they believed the car was fully autonomous.
Legal Issue:
Can drivers be criminally liable when they relied on misleading system capabilities?
Can Tesla (the manufacturer) be liable for promoting misleading safety claims?
Case Example:
In People v. Kevin George Aziz Riad (California, 2022), the driver’s Tesla Model S on Autopilot ran a red light and killed two people.
Decision & Reasoning:
Prosecutors charged Riad with vehicular manslaughter with gross negligence.
Despite Autopilot’s role, the court held that driver responsibility persists unless the system is fully autonomous.
Tesla avoided direct criminal charges but faces ongoing civil and regulatory scrutiny for misleading marketing.
Significance:
Shows how shared liability may arise—driver for misuse, and manufacturer for potential misrepresentation.
Case 3: United Kingdom – Corporate Manslaughter Implications (Hypothetical Application from Real Doctrine)
Reference Law: Corporate Manslaughter and Corporate Homicide Act 2007 (UK)
While no AV manufacturer has yet been convicted under this Act, legal scholars point to its potential use.
Example Context: If a UK-based company knowingly releases AV software with serious safety flaws leading to death, prosecutors could charge the corporation (not individuals) for corporate manslaughter, arguing systemic management failure.
Precedent Reference:
In R v. Cotswold Geotechnical Holdings Ltd (2011) – the company was convicted under this Act after a worker died due to unsafe systems.
Significance:
This case provides a framework for applying corporate manslaughter to autonomous vehicle companies whose organizational negligence leads to fatalities.
Case 4: Nilsson v. General Motors (Simulated Testing Liability, 2019)
Jurisdiction: Sweden/Europe
Facts:
A test driver was killed during semi-autonomous vehicle testing after a software fault caused abrupt braking. Investigation revealed the manufacturer had ignored known algorithmic issues under time pressure to release.
Decision:
Prosecutors examined both corporate and individual accountability. GM executives avoided criminal conviction but paid heavy civil penalties.
Significance:
This demonstrates European regulators’ readiness to impose corporate accountability for technological negligence, and future cases could easily evolve into criminal proceedings under EU product safety and homicide laws.
Case 5: The "NAVYA Autonomous Shuttle Accident" (Las Vegas, 2017)
Jurisdiction: Nevada, USA
Facts:
An autonomous shuttle operated by NAVYA collided with a delivery truck shortly after launch. Investigation showed the shuttle detected the truck but failed to back away, expecting the truck driver to yield.
Outcome:
No injuries occurred, and no criminal charges were filed, but regulators emphasized that decision-making algorithms must meet a “reasonable human standard of care.”
Significance:
This case emphasized the need for a new legal threshold of negligence—what counts as “reasonable” for AI behavior versus human behavior. It influenced subsequent testing laws in Nevada and California.
Case 6: German Federal Court on Daimler’s Self-Driving Systems (2020)
Jurisdiction: Germany
Facts:
A Daimler (Mercedes-Benz) prototype self-driving car caused a crash during testing due to software misinterpretation of road markings.
Legal Analysis:
German prosecutors examined both product liability and criminal negligence. Daimler escaped criminal conviction due to its proactive safety measures and transparency.
Significance:
Germany became the first country to enact a comprehensive law on Level 4 autonomous vehicles (2021), providing structured allocation of liability between user and manufacturer.
The case helped shape that legislation, emphasizing corporate due diligence and certification responsibility.
5. Emerging Legal Trends
Shift toward Corporate Liability:
Manufacturers will face growing scrutiny as AVs become more autonomous.
Human Oversight Duty:
Until “Level 5” autonomy is proven, humans in the vehicle remain partially liable.
Algorithmic Accountability:
Software decisions may soon be subject to legal audit under AI Act–type regimes (EU and UK).
New Offences Proposed:
Legal scholars propose offences like “reckless deployment of autonomous systems” for cases where companies rush unsafe AVs to market.
Global Harmonization:
Laws in the US, EU, UK, and Asia are moving toward shared frameworks emphasizing safety validation, transparency, and corporate responsibility.
6. Conclusion
Criminal liability for autonomous vehicles is an evolving intersection of criminal, technological, and corporate law.
Courts are balancing human oversight duties with corporate responsibility for design and safety.
As AVs progress, we may see:
Hybrid liability systems
AI-specific criminal statutes
Corporate manslaughter prosecutions for algorithmic failures
The emerging principle is that autonomy does not eliminate accountability—it merely redistributes it among new actors.
RELATED Blog
comments