By: Counselor Amr el salte | Lawyer before the Court of Cassation and the Supreme Administrative Court

Presumed Criminal Liability for Out-of-Control Digital Entities: A Comparative Study of Criminal, Civil, and Administrative Liability

The rapid advancement of artificial intelligence (AI) technologies and autonomous digital systems has enabled digital entities to make decisions and execute actions without direct human intervention. This unprecedented development has raised significant legal questions regarding liability for the damages that may result when such entities operate beyond human control.

This issue gives rise to a fundamental legal question: Who bears responsibility when an autonomous digital entity commits an unlawful act, a criminal offense, or causes substantial harm? Should liability remain exclusively attributed to human beings, or does technological progress require the development of new legal frameworks capable of addressing the challenges of the digital revolution?

This study examines the presumed criminal, civil, and administrative liability arising from damages caused by out-of-control digital entities while highlighting the most significant comparative legal approaches.

First: Criminal Liability

Under traditional criminal law, criminal liability is established upon the existence of two essential elements:

• The material element (actus reus).

• The mental element (mens rea), whether criminal intent or negligence.

Accordingly, criminal liability may only be imposed upon a natural person or a legal entity recognized by law.

A digital entity or artificial intelligence system, however, does not currently possess an independent criminal legal personality. Consequently, it cannot be prosecuted as the principal offender in a criminal act.

The legal dilemma arises when an AI system makes an unforeseen decision that neither its developer nor its user anticipated, resulting in serious harm.

Examples include:

• A self-driving vehicle causing a fatal accident.

• A surgical robot committing a fatal medical error.

• An AI-powered financial system triggering a severe economic collapse.

• An autonomous digital platform conducting fraudulent transactions without direct human involvement.

In such cases, the digital system itself cannot be punished. Instead, the law seeks to identify the legally responsible human or corporate party.

Who Bears Criminal Liability?

The answer depends on the circumstances that caused the digital system to operate beyond control.

Criminal liability may fall upon:

• The software developer.

• The manufacturer.

• The technology company.

• The owner of the system.

• The end user.

• The operating organization.

This liability may arise where there is evidence of:

• Negligence.

• Professional misconduct.

• Prior knowledge of foreseeable risks.

• Failure to update the system.

• Inadequate cybersecurity or safety measures.

• Allowing continued operation despite the discovery of serious defects.

Can Criminal Liability Be Presumed?

A growing body of legal scholarship argues that the unprecedented evolution of artificial intelligence may justify introducing a new legal concept referred to as "Presumed Criminal Liability."

Under this approach, the producer or operator of a digital system may be presumed criminally liable whenever damage results directly from that system, unless they successfully establish the existence of an external cause, force majeure, or unlawful third-party interference.

Nevertheless, this approach remains highly controversial because presumed criminal liability may conflict with fundamental criminal law principles, particularly the principles of personal criminal responsibility and the presumption of innocence.

Comparative Legal Approaches

European Union

The European Union does not recognize artificial intelligence as possessing an independent criminal legal personality. Instead, it imposes extensive legal obligations upon developers and operators of high-risk AI systems, including strict requirements concerning risk management, transparency, human oversight, and accountability.

United States

Most American jurisdictions continue to apply traditional liability rules by assigning criminal responsibility to the individual or organization responsible for negligent design, programming, operation, or supervision, without recognizing AI as an independently liable criminal subject.

Japan and South Korea

Both jurisdictions focus primarily on professional liability and corporate responsibility while developing specialized regulatory standards governing robotics and intelligent systems without granting them independent legal personality.

Second: Civil Liability

Civil liability is considerably more flexible than criminal liability, as compensation generally depends upon proving the occurrence of damage and establishing a causal relationship between the conduct and the resulting harm.

Civil liability may arise from:

• Fault-based liability.

• Liability for things.

• Custodian liability.

• Product liability.

• Strict liability under certain legal systems.

Liability of Manufacturing Companies

Where a digital system causes damage due to:

• Programming defects.

• Design flaws.

• Machine learning failures.

• Weak cybersecurity safeguards.

The manufacturing or developing company may bear full civil liability for compensating individuals or institutions that suffered damage.

Liability of the User

A user may also incur civil liability where they:

• Ignore safety warnings.

• Operate the system improperly.

• Disable security mechanisms.

• Input inaccurate data resulting in damage.

Third: Administrative Liability

Regulatory authorities bear responsibility for supervising, regulating, and monitoring the use of digital systems within their respective jurisdictions.

Administrative sanctions may include:

• Financial penalties.

• Suspension of activities.

• Revocation or suspension of licenses.

• Withdrawal of operational permits.

• Mandatory system modifications.

• Imposition of enhanced cybersecurity requirements.

In highly sensitive sectors such as healthcare, transportation, and energy, regulatory oversight may be even more stringent to safeguard the public interest.

Legal Challenges

Current legal systems face several significant challenges, including:

• The absence of a unified legal definition of autonomous digital entities.

• Difficulties in identifying the legally responsible party where multiple stakeholders are involved.

• The capacity of AI systems to evolve through self-learning after deployment.

• Challenges in establishing causation between programming decisions and resulting harm.

• The potential conflict between presumed liability and established principles of criminal justice.

Do We Need New Legislation?

A considerable number of legal scholars believe that traditional legal principles are no longer sufficient to regulate damages arising from artificial intelligence.

Accordingly, modern legislation should address the following issues:

• Clearly defining the legal responsibilities of every party involved in the development, deployment, and operation of intelligent systems.

• Establishing mandatory standards for transparency and explainability of AI-generated decisions.

• Requiring compulsory insurance for high-risk AI systems.

• Creating specialized evidentiary rules governing crimes and damages caused by artificial intelligence.

• Developing comprehensive regulatory mechanisms for supervision, risk disclosure, and cybersecurity updates.

Conclusion

Artificial intelligence and autonomous digital entities have become an undeniable reality that challenges legal systems worldwide. While the prevailing legal principle continues to hold that criminal liability may only be imposed upon natural persons or legally recognized entities, the rapid pace of technological advancement necessitates reconsidering traditional doctrines of legal responsibility, particularly where autonomous digital systems operate beyond human control and cause serious harm.

Developing a balanced legislative framework that protects individuals, encourages technological innovation, and clearly allocates legal responsibility represents one of the most significant challenges facing contemporary legal systems in the age of artificial intelligence.

---

About Counselor Amr El Salt

Counselor Amr El Salt
Attorney before the Egyptian Court of Cassation and the Supreme Administrative Court

For Legal Consultations and Appointments:

📱 WhatsApp: +20 1223424679

☎️ Phone: +20 1012849603

✉️ Email: [email protected]