1. Introduction

Responsibility attribution in hybrid systems refers to the legal process of identifying who should be held accountable when harm, negligence, unlawful conduct, or regulatory violations arise from the combined activities of human beings, artificial intelligence (AI), automated technologies, biological systems, software, machines, and institutional decision-makers.

A hybrid system is a system in which multiple human and non-human components interact to perform a particular function. Examples include:

AI-assisted medical diagnosis and treatment.

Human-controlled autonomous vehicles.

Biotechnology laboratories using automated genetic-editing systems.

Human–AI decision-making in clinical trials.

Robotic surgery and automated pharmaceutical manufacturing.

Biological monitoring systems connected to predictive algorithms.

Hybrid research environments combining human expertise, AI models and automated laboratory equipment.

The central legal question is:

When a hybrid system causes harm, who is legally responsible: the human operator, developer, manufacturer, researcher, institution, regulator, or the organisation controlling the system?

Traditional legal systems generally assign responsibility to identifiable persons or organisations. Hybrid systems complicate this approach because decisions and outcomes may emerge from the combined actions of several participants rather than from one individual.

Responsibility attribution therefore requires a framework that considers human involvement, technical control, foreseeability, causation, negligence, institutional supervision and regulatory compliance.

2. Meaning and Nature of Hybrid-System Responsibility

Responsibility attribution in hybrid systems involves identifying the legal consequences of the interaction between different actors and technological components.

Unlike conventional systems, where one person may directly perform an act causing injury, hybrid systems distribute functions across several participants.

For example, consider an AI-assisted biotechnology laboratory:

A researcher selects a genetic-editing objective.

An AI system recommends a genetic modification.

Automated laboratory equipment performs the experiment.

A biotechnology company supplies the software and equipment.

A research institution supervises the project.

A regulatory authority establishes safety requirements.

If the genetic intervention causes unexpected biological harm, several questions arise:

Was the researcher negligent?

Was the AI software defective?

Did the manufacturer fail to provide adequate safeguards?

Did the institution fail to supervise the experiment?

Were regulatory requirements ignored?

Did the automated system behave unpredictably despite appropriate precautions?

The law must distinguish between technical participation and legally relevant responsibility.

An important principle is that participation in a hybrid system does not automatically establish liability. Responsibility must be established through applicable legal duties, statutory provisions, causation, fault or strict-liability principles.

3. Major Theories of Responsibility Attribution

A. Direct Responsibility

Direct responsibility arises when a person personally performs an act or omission that causes legally recognised harm.

For example, a laboratory scientist may knowingly override a mandatory biosafety mechanism and expose workers to a dangerous biological agent.

The scientist may face liability if the necessary elements of negligence or another applicable legal offence are established.

Direct responsibility may arise against:

Researchers.

Laboratory technicians.

Medical practitioners.

System operators.

Supervisors.

Institutional administrators.

The important question is whether the individual's conduct breached an applicable legal obligation.

B. Vicarious Responsibility

Vicarious liability arises when an employer or organisation is held legally responsible for wrongful acts committed by an employee in the course of employment.

For example, an employee operating an automated pharmaceutical manufacturing system may negligently disregard safety procedures, resulting in contaminated products.

Depending on the applicable law and employment relationship, the employer may be liable alongside the employee.

This doctrine is important because organisations control staffing, training, supervision and operational procedures.

However, an employer is not automatically liable for every independent act of an employee. The connection between the wrongful conduct and the employment must be established.

C. Product-Based Responsibility

Product liability concerns harm caused by defective products.

In hybrid systems, defective components may include:

AI-controlled medical devices.

Automated laboratory equipment.

Genetic-testing instruments.

Robotic surgical systems.

Biotechnology manufacturing software.

Biological monitoring devices.

A manufacturer may be responsible where a legally recognised product defect, inadequate warning or manufacturing failure causes harm.

In India, the Consumer Protection Act, 2019 provides a statutory product-liability framework under Sections 82–87.

Responsibility may extend to manufacturers, product service providers and sellers, subject to the applicable statutory conditions.

D. Institutional Responsibility

Institutional responsibility arises when an organisation fails to establish or maintain adequate systems of governance and supervision.

Examples include:

Failure to conduct risk assessments.

Inadequate laboratory safety procedures.

Failure to investigate technical malfunctions.

Insufficient employee training.

Failure to maintain system audit records.

Negligent approval of high-risk experiments.

A research institution cannot necessarily avoid responsibility merely by arguing that an automated system independently produced the harmful outcome.

Its own management failures must be examined.

E. Shared or Distributed Responsibility

Distributed responsibility is particularly important where multiple participants contribute to a harmful outcome.

For example, a robotic surgical accident may involve:

A defective software algorithm.

An inadequate manufacturer warning.

Incorrect hospital configuration.

Insufficient medical supervision.

Failure by the surgeon to respond to an emergency.

Depending on the facts and applicable law, more than one party may be legally responsible.

The court must distinguish individual contributions, legal duties and causal connections instead of automatically assigning all responsibility to one participant.

4. Essential Legal Principles for Attribution

These principles provide the foundation for allocating responsibility in complex technological environments.

5. Important Case Laws

The following decisions are particularly relevant. Some directly establish general principles of liability, while others provide analogous principles that can be applied to emerging hybrid technologies. They should not be mistaken for judgments specifically deciding AI or biotechnology liability.

Case 1: M.C. Mehta v. Union of India (1987) 1 SCC 395

Important Indian precedent

Facts:

The case concerned the leakage of oleum gas from the Shriram Foods and Fertiliser Industries in Delhi. The incident raised serious questions concerning industrial safety, public health and the responsibility of enterprises operating hazardous activities.

Legal issue:

Whether an enterprise engaged in hazardous or inherently dangerous activities could escape responsibility by arguing that it had exercised reasonable care.

Judgment:

The Supreme Court of India developed the principle of absolute liability.

An enterprise engaged in hazardous or inherently dangerous activities owes an absolute and non-delegable duty to ensure that no harm results from those activities.

If harm occurs because of the hazardous activity, the enterprise must compensate affected persons, without relying on the traditional exceptions associated with the rule in Rylands v. Fletcher.

Relevance to hybrid systems:

This decision is particularly relevant to biotechnology laboratories, synthetic biology facilities, genetic engineering operations and automated hazardous manufacturing systems.

An organisation cannot simply transfer its safety obligations to an automated machine or external technology provider.

The enterprise remains accountable under the applicable absolute-liability principle where its hazardous activity causes harm.

Case 2: Donoghue v. Stevenson [1932] AC 562

Facts:

A consumer became ill after allegedly consuming ginger beer containing a decomposed snail. The manufacturer had supplied the product in a sealed bottle.

Legal issue:

Whether a manufacturer owed a duty of care to the ultimate consumer despite the absence of a direct contractual relationship.

Judgment:

The House of Lords recognised the modern neighbour principle, establishing that a person must take reasonable care to avoid acts or omissions that could foreseeably injure persons closely and directly affected by their conduct.

Relevance to hybrid systems:

Manufacturers of AI-controlled equipment, genetic-testing devices and automated medical technologies may owe duties of care to foreseeable users and affected persons.

The absence of a direct contract between a technology manufacturer and an injured patient does not necessarily eliminate the possibility of a negligence claim.

Case 3: Caparo Industries plc v. Dickman [1990] 2 AC 605

Facts:

Caparo Industries relied on audited financial statements when acquiring shares in a company. It subsequently alleged negligence by the auditors.

Legal issue:

Under what circumstances does a duty of care arise in negligence?

Judgment:

The House of Lords articulated three important considerations:

Reasonable foreseeability of harm.

Proximity between the parties.

Whether imposing a duty is fair, just and reasonable.

Relevance to hybrid systems:

This framework assists courts in determining whether a developer, software provider, institution or operator owes a duty of care to a person injured by a hybrid system.

For example, a software developer supplying clinical decision-support technology may owe a duty in circumstances where the relevant legal requirements of duty are satisfied.

The case is especially important when responsibility must be allocated among actors who have no direct contractual relationship.

Case 4: Jacob Mathew v. State of Punjab (2005) 6 SCC 1

Facts:

The case involved allegations of medical negligence arising from the treatment of a patient who died after experiencing breathing difficulties.

Legal issue:

What standard should be applied when determining professional negligence, particularly in relation to medical practitioners?

Judgment:

The Supreme Court explained that professional negligence must be evaluated against the standard of reasonable competence and care expected from a professional in the relevant circumstances.

A mere error of judgment does not automatically establish negligence.

Relevance to hybrid systems:

This decision is relevant to AI-assisted medicine and robotic surgery.

A medical practitioner who uses an AI-generated recommendation is not automatically exempt from responsibility.

The court would need to examine whether the professional acted with reasonable competence, whether reliance on the system was justified and whether the professional's conduct caused the injury.

Case 5: Rylands v. Fletcher (1868) LR 3 HL 330

Facts:

The defendants constructed a reservoir on their land. Water escaped through old mine workings and flooded the claimant's neighbouring mine.

Legal issue:

Whether a person who brings and keeps a potentially dangerous substance on their land can be liable when it escapes and causes damage.

Judgment:

The House of Lords established the traditional rule of strict liability for certain non-natural uses of land involving the escape of dangerous things.

Relevance to hybrid systems:

The principle is historically important when examining dangerous biological materials, hazardous laboratory substances and industrial systems.

However, Indian law developed a stronger absolute-liability doctrine for hazardous enterprises in M.C. Mehta v. Union of India.

Case 6: Bolitho v. City and Hackney Health Authority [1998] AC 232

Facts:

A child suffered severe injury after a medical emergency. The case examined whether a doctor's failure to attend and intubate the child amounted to negligence.

Legal issue:

Whether professional medical opinion alone is sufficient to establish that a particular course of treatment is reasonable.

Judgment:

The House of Lords held that professional opinion relied upon to establish the appropriate standard of care must withstand logical analysis.

Relevance to hybrid systems:

An institution cannot automatically defend an AI-assisted clinical decision merely by claiming that its software recommendation was accepted industry practice.

The safety and rationality of the system's use must be examined in the circumstances.

Case 7: State of Rajasthan v. Vidyawati AIR 1962 SC 933

Facts:

A government-employed driver negligently caused an accident while driving a government vehicle.

Legal issue:

Whether the State could be held responsible for the negligent acts of its employee.

Judgment:

The Supreme Court recognised the possibility of State vicarious liability for tortious acts committed by employees in the course of employment.

Relevance to hybrid systems:

The case provides an important analogy for determining institutional and employer responsibility when employees operate automated systems.

A research institution or organisation may face liability for negligent acts of employees operating hybrid technologies, depending on the applicable legal framework.

Case 8: NTSB Investigation of the Uber Automated Vehicle Crash (2019)

Facts:

In March 2018, an Uber developmental automated-driving vehicle struck and killed a pedestrian in Tempe, Arizona.

The vehicle was operating in autonomous mode, with a human safety operator present.

Investigation findings:

The US National Transportation Safety Board identified the operator's failure to monitor the driving environment as the probable cause.

Contributing factors included:

Inadequate safety risk assessment by Uber.

Ineffective oversight of vehicle operators.

Insufficient mechanisms addressing automation complacency.

Inadequate safety culture.

Insufficient state oversight of autonomous-vehicle testing.

Relevance to hybrid systems:

This investigation provides an important practical example of distributed responsibility.

It demonstrates that responsibility analysis must examine both the human operator and the organisation designing and supervising the automated system.

Legal significance: This was an accident investigation, not a judicial precedent determining civil liability.

NTSB

6. Responsibility Attribution in Biotechnology and AI-Hybrid Systems

This area is particularly important because biological systems may interact with automated technologies in ways that create risks extending beyond the original research environment.

Consider the following examples.

A. AI-assisted genetic engineering

Suppose an AI system recommends a genetic modification that produces an unintended biological effect.

Potentially responsible parties include:

The researcher who approves the intervention.

The developer responsible for defective software.

The equipment manufacturer.

The laboratory responsible for safety supervision.

The institution responsible for research governance.

The court should examine the contribution of each participant, including whether the harm was foreseeable and preventable.

B. Automated clinical biotechnology

Suppose an automated diagnostic system incorrectly identifies a genetic disorder and recommends inappropriate treatment.

Potential responsibility may arise from:

Defective algorithm design.

Inadequate clinical validation.

Negligent interpretation of results.

Failure to disclose known limitations.

Inadequate institutional supervision.

Medical negligence, consumer protection and product-liability principles may become relevant.

C. Synthetic biology and autonomous laboratories

Autonomous laboratories combine robotic equipment, software, biological samples and human research supervision.

A malfunction may result in contamination, unauthorised experimentation or exposure to dangerous biological materials.

Responsibility may arise from defective equipment, inadequate containment, negligent supervision or failure to comply with biosafety requirements.

D. Human–machine biological interfaces

Hybrid biological interfaces may involve neural devices, implanted sensors, automated drug-delivery systems and brain-computer interfaces.

If a device generates an unintended biological response, responsibility may depend on:

Design defects.

Software malfunction.

Inadequate clinical testing.

Failure to monitor adverse effects.

Improper medical intervention.

Lack of informed consent.

The legal framework must consider the respective obligations of manufacturers, clinicians, researchers and healthcare institutions.

7. Indian Legal Framework

India does not currently have a comprehensive, standalone liability statute specifically governing responsibility attribution in AI-hybrid systems.

Existing legal principles and legislation provide the principal framework.

Two important distinctions must be maintained:

An AI system is not automatically treated as an independent legal person capable of bearing ordinary civil or criminal liability.

An organisation cannot automatically escape its own legal obligations simply because a technological component participated in the harmful activity.

For AI-related claims, negligence, product liability, contractual liability and other applicable statutory doctrines must be assessed according to the facts.

Chambers and Partners+1

8. International Legal Developments

A. European Union

The European Parliament has examined the difficulties of assigning liability when AI systems operate autonomously.

Its 2020 resolution on civil liability for AI considered the responsibility of operators and producers, including situations in which autonomous systems cause harm.

The European Parliament also adopted a proposed AI liability framework in 2020, but that resolution should not be confused with an enacted EU-wide AI civil-liability regulation.

The central policy concern is that victims should not lose access to compensation merely because the technological process is difficult to understand.

EUR-Lex+1

B. United States

US liability principles generally rely on established legal doctrines, including:

Negligence.

Vicarious liability.

Strict product liability.

Manufacturing defects.

Design defects.

Failure to warn.

The responsible party may depend on whether the harm resulted from defective technology, negligent operation or inadequate organisational supervision.

C. United Kingdom

UK courts use established negligence and product-liability principles to examine the conduct of manufacturers, professionals and organisations.

The decisions in Donoghue v. Stevenson, Caparo v. Dickman and Bolitho v. City and Hackney Health Authority provide useful foundations for analysing duties of care and professional responsibility.

9. Major Challenges in Responsibility Attribution

1. The responsibility gap

An autonomous system may produce harmful outcomes without a clearly identifiable human decision immediately preceding the event.

2. Algorithmic opacity

It may be difficult to establish why a machine generated a particular output.

3. Multiple-party involvement

Several actors may contribute to the same injury, making individual responsibility difficult to establish.

4. Evidentiary difficulties

Important evidence may remain within the control of software developers, manufacturers or research institutions.

5. Unpredictable system behaviour

Self-learning technologies may produce outcomes that were not anticipated during initial testing.

6. Cross-border responsibility

Developers, manufacturers, laboratories and injured persons may be located in different countries, creating jurisdictional and choice-of-law complications.

7. Inadequate regulatory supervision

Regulatory authorities may lack the technical capacity to investigate complex hybrid-system failures.

These challenges are particularly significant in autonomous laboratories, genetic engineering, AI-assisted medicine and advanced biological research.

EUR-Lex+2

10. Proposed Legal Framework for Responsibility Attribution

A comprehensive responsibility-attribution framework should contain the following safeguards:

A particularly important safeguard is the creation of an auditable responsibility chain.

Every important decision should be traceable to the relevant human participant, technical component, organisation or regulatory process.

However, traceability alone does not establish liability. It supplies evidence for determining whether the relevant legal requirements have been satisfied.

11. Critical Legal Analysis

Traditional legal responsibility is based largely on human conduct, organisational duties, causation and legally recognised risks.

Hybrid systems challenge these assumptions because a harmful outcome may emerge from interactions between components controlled by different participants.

Three competing approaches can be identified:

Individual responsibility: Focuses on the person whose act or omission directly contributed to the harm.

Organisational responsibility: Focuses on the institution that designed, deployed, supervised or controlled the system.

Distributed responsibility: Examines the legally relevant contributions of multiple actors and allocates liability according to applicable legal principles.

The third approach is particularly useful for understanding complex hybrid systems, but it must not become a means of diluting individual accountability.

For example, a manufacturer should not escape responsibility for a defective safety mechanism merely because a researcher operated the equipment. Similarly, a researcher should not automatically escape responsibility for deliberately overriding safety controls by blaming the manufacturer.

The law must therefore distinguish between technical causation, legal causation and the allocation of responsibility.

12. Conclusion

Responsibility attribution in hybrid systems represents an important development in contemporary legal jurisprudence.

As AI, biotechnology, robotics and automated decision-making increasingly interact, traditional legal concepts must be applied to complex networks of human and technological participation.

The decisions in M.C. Mehta v. Union of India, Donoghue v. Stevenson, Caparo Industries v. Dickman, Jacob Mathew v. State of Punjab and Bolitho v. City and Hackney Health Authority provide important foundations for determining responsibility.

The central legal principle is that technological autonomy should not automatically eliminate human or institutional accountability.

A fair responsibility framework must identify the relevant duty, examine the conduct of each participant, establish causation and apply the appropriate liability standard.

Ultimately, the objective should be to ensure that technological complexity does not become a legal escape route for those who create, manufacture, operate, supervise or control potentially harmful hybrid systems.

13. Important Cases and Research References

Further research materials:

Supreme Court judgment – M.C. Mehta v. Union of India.

Caparo Industries v. Dickman.

Where Does Responsibility Lie? – Medical Law Review.

European Parliament study – Artificial Intelligence and Civil Liability.

NTSB – Uber Automated Vehicle Crash Investigation.

Academic note: The above analysis distinguishes established judicial precedents from regulatory materials and analogous technology investigations. There is not yet a settled body of case law specifically allocating civil liability for harm caused by fully autonomous biotechnology–AI hybrid systems.