Corporate Liability for AI Harms and Monopolistic Practices

Introduction

Artificial intelligence is no longer a far-off fantasy, dominating approval of loans, diagnosis, employment and even on the roads with autonomous vehicles. However, when the AI systems malfunction or discriminate against persons, the big question is, who is held legally liable. By referring to an algorithm, corporations cannot get away with liability. A company as a legal entity is responsible under English law, by human or machine decisions, to be accountable as a legal entity. This blog reviews how current and future laws ascribe damages caused by AI to companies and how AI market concentration presents an opportunity to invoke antitrust. The main argument is easy to grasp: automation increases accountability and the legislation is rapidly diverging in the right direction, both in tort and competition.

Corporate Liability of AI Injuries

Most jurisdictions apply the same doctrines of tort negligence, strict liability and product fault to AI damages. Under the Restatement (Third) of Torts: Products Liability § 2(b) (1998), manufacturers and operators are required to guard against foreseeable harm. In an instance where an AI system hurts an individual in a manner that ought to have been foreseen by a responsible developer, the corporation might be held responsible for shoddy design, lack of adequate testing, or improper supervision. Theories of product liability such as design flaw and lack of warning are being applied to AI. An unethical AI medical system making incorrect diagnosis of patients, such as one, can result in a negligence claim as well as a failure-to-warn claim when the developer was aware of the known limitations.

One notable case is Moffatt v. Air Canada (2024) when the British Columbia Civil Resolution Tribunal ruled that the airline was vicariously liable as a result of false statements made by its AI chatbot. This decision is clear cut: it is the corporation, and not the secret software that is at fault. The Law Commission of England and Wales (2022) has accepted this principle as it precludes those arguments according to which the liability can be out-contracted on machines.

This has been greatly solidified by European law. The EU AI Act (2024) requires high-risk AI systems, such as employment, credit scoring, and other essential services, to meet traceability requirements, human oversight, and stringent risk management. The new Product Liability Directive (2024) extends this to software and even AI more generally: a company can be found guilty of physical or psychological harm caused by its AI product, without having to prove fault, and liabilities are required to be disclosed when defences are heightened.

The negligence and strict liability two-pole discussion is still ongoing. Anat Lior, who is a scholar, proposes strict liability of harms of AI, claiming that the standards of negligence would be inappropriate when it comes to unexpected malfunctions of the algorithms. Those opposing it argue that strict liability unfairly weighs heavily on start-ups. Lior replies that the monopolisation of AI is already a process and strict liability can be balanced by well-tuned insurance mechanisms. The general flow, nevertheless, is evident: the companies that use AI will either need to internalize risk management, bias auditing, and people supervision, or have to face legal penalties.

Monopolistic Behaviour of the AI Business

The concentration of the AI industry is noticeable. The key inputs to AI are mostly controlled by only a few companies including Google, Microsoft, Amazon, Meta and NVIDIA: proprietary data, customised chips (GPUs), cloud technology and user-facing interfaces. Such a strong concentration creates strong network effects: the better AI works, the more people are willing to use it, and they in turn increase AI performance;, consolidating market power and creating entry barriers. According to the European Commission (2024), the danger with this type of feedback loop is that the created platforms can act as gatekeepers, with the control on the use of AI tools extending to whole markets.

There are a number of monopolistic risks identified by regulators. First, leading companies can pursue exclusionary practices, such as integrating AI software with proprietary hardware, giving their offerings self-preference, or obliging one-way contracts to shut out competitors. An example of this risk is Google’s long-standing practice of integrating its search engine with the Android operating system on mobile devices, a practice that has drawn scrutiny from the European Commission. In the United States v. Google LLC (2024) case, a federal court in the United States determined that Google had unlawfully maintained its monopoly in search through exclusive dealing arrangements and ordered remedies accordingly. These remedies are now actively applied to the AI-based services that Google provides, in recognition that AI represents a new phase of dominance in the market.

Second, algorithmic collusion is becoming a new separate antitrust threat. State of the art pricing algorithms can align market behaviour without any communication involving a human being, bypassing the classic cartel law. This move by the DOJ against RealPage (2024-2025), in which landlords were accused of using a joint algorithm to set prices, killing competition, which led to a consent decree requiring continued compliance monitoring. California has already acted legislatively: the common algorithmic pricing tool is outlawed under AB 325 (2025), which addresses the situation in competition between businesses.

Third, massive investments and collaborations in AI - USD 13 billion investment between Microsoft and OpenAI, investments by Amazon in spaces like Anthropic, and the OpenAI-Oracle-SoftBank Stargate project have the potential to become pipeline pollutants even without initiating a formal merger investigation. The FTC and DOJ Joint Statement on Generative AI (June 2024) indicates that there will be a greater emphasis on agencies on such alliances irrespective of their structural nature and whether it entails the existence of bottlenecks to AI data and infrastructure.

Competition Law and Liability Law

Tort law and antitrust law relate to two different fronts of AI responsibility, of paying victims of harm caused by AI, and the other upholding the competitive process. Nevertheless, they go hand in hand. Breaking a monopoly in search does not just unleash markets to the forces of competition; it also ensures that no single company will again reap the scale benefits of unchecked algorithms. Conversely, capital requirements like transparency, compulsory disclosure of risks, and auditability may initiate patterns of harmful conduct in industries noticeable to rivals enforcers.

This convergence means that corporate governance is obligated by new responsibilities. Any given duty of reasonable care of directors of companies that adopt AI gives rise to the duty of AI supervision which is a duty that the Companies Act 2006, of the UK also grants. Whether bent humanly or algorithmically the choices taken by computers are under section 174 of the Companies Act 2006. AI risk management must be embraced as corporate strategy, not as a cost of regulation but as a governance need by the board. This is reinforced by the EU Digital Markets Act (2022) that makes self-preferencing and tying practices by recognised gatekeeper platforms illegal, signalling that the scope of the DMA could be broadened to include new AI platforms.

Conslusion

In the AI age, equal responsibility should be taken in both tort and competition law. Long-established principles of corporate liability related to materials-based products are undergoing transformation with intent to AI, and novel regulatory instruments must make strict regulative responsibility of AI suppliers a reality in the next few years. At the same time antitrust regulators in the U.S., the EU, and the UK are grappling with the fact that just a few companies, without any regulation, can become embedded in the control of the architecture of the AI economy.

It is clear that there is no such message as automation immunises a corporation against liability, or innovation warrants a market monopoly. Corporations have to construct security forces, equitable AI systems, and rival in open markets. The competition and liability law are adapted to each other to ensure that no technological advancement is coupled with a disregard for responsibility, consumer interests, or the basic rights of the people the rule of law is meant to protect. The legislation is mobility-, and companies must follow suit.

---

Co-authored by Prachi Amit Sharma and Deepika Singh, fifth-year students of School of Law,Christ (Deemed to be University).