LawTechnos | Legal Technology Magazine

At the heart of the global technology trends is the buzz surrounding the subject of artificial intelligence. The ability of machines to think and act like humans has been a major interrogation undertaken by computer scientists like John McCarthy who is today known as the father of AI. 

Today, AI machines are used in several industries – manufacturing, legal, finance, health, to name a few. The underlying merit is that they perform faster, are less prone to mistakes and are efficient. However, with the growing use of these machines, it has become pertinent for laws to evolve with. Essentially, to regulate its usage and answer certain legal questions such as liability, privacy and data protection, etc. It is important to state that these laws are not geared at hampering the development of AI, but to make it civil, transparent and suitable for human enjoyment. Examples of such laws include General AI Laws, Data Protection Laws, Intellectual Property Laws, Criminal & Civil Laws (the most of which are currently undergoing an overhaul to reflect the changes in technology), etc.

What is Artificial Intelligence?

Human or natural intelligence is that conceived naturally. On the flip side, an intelligence is said to be artificial where it is configured or programmed by other means other than by nature. In essence, when a thing or an object (inanimate) such as a machine is attributed with an intelligence to act or carryout tasks like an animate, such intelligence is considered artificial. 

AI refers to the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions. According to John McCarthy, it is the science and engineering of making intelligent machines. These machines are either programmed to carryout specialized tasks like in the case of Chatbots or perform general tasks. In the latter case, such machines learn from experience and can independently analyze data from a set of behaviors and arrive at a conclusion. 

AI and The Law: The Web

The nexus between AI and Law is to some extent blurry. This assertion is true because while AI is a subject in the computer sciences, Law is a subject in humanities, and whereas computer science has and will continue to experience speedy evolution, Law on the other hand, has been a conservative discipline – sluggish to changes. As such, rather than being proactive, it is always trying to catch up with the changes. 

A relationship between AI and Law can however be drawn from recent attempts by the legislators to regulate the usage and application of AI. At the core of these are laws that protect human privacy and data utilized by AI machines, product liability provisions contained in consumer protection laws as well as, intellectual property laws.

1. Data Protection:  AI machines utilize data to arrive at conclusions. An AI-enhanced camera in determining whether a person is a threat or a potential threat will examine the person (face, pockets, behaviors etc) before raising an alarm. In such circumstances, the personal data of that person is collected and in most cases stored. In the event of a hack, the stored data is utilized by the hacker to perpetrate harmful acts against the data subject.

In most cases, the person whose data is collected by the AI is not informed of whether the machine stores his data and for what duration. All of these constitute wrong practices. First, the protection of one’s data is not guaranteed. Secondly, the subject is not informed and his consent is not sought. It is against this backdrop that laws on data protection have emerged, with internationally recognized principles such as consent, transparency, security, purpose limitation, amongst others.  The most comprehensive being the General Data Protection Regulation (GDPR) in the EU, and others including, The Personal Data Protection Law in the UAE, the Nigeria Data Protection Regulation (NDPR), and a host of others. 

2. Product Liability: Product liability is the liability incurred by a manufacturer for producing a defective product. This doctrine permits a consumer or user of a defective or misrepresented product to bring an action against the producer. A product is defective if it does not meet the safety requirements which a person could reasonably expect. A product liability case may arise from contract, tort (under negligence or strict liability) or by statute.

In context, in the event that a driverless car fails to avoid an accident that later analysis shows was avoidable or an AI-enabled robotic tool causes an avoidable harm to a patient during a surgery, who is to be held liable – the machine or its producer? This is what is known as the question of attribution. Product liability in the AI context will include negligence, manufacturing defects, failure to warn, misrepresentation and breach of warranty. Under negligence for example, manufacturers are obligated to make products that when used in a reasonably foreseeable way will not cause harm. If, however, an AI product when used in a reasonably foreseeable way yet causes harm, the manufacturer may be liable under negligence. In essence, programmers or producers of an AI tool are generally liable for any defect. 

A concern however, is the ability of AI machines to learn and produce algorithms on their own. In such cases where the algorithm designed largely by the AI makes a mistake, who is to be liable?  Although this question may seem impossible to answer, the liability still rests on the producer of the AI not the AI itself. Even though AI enables learning (popularly known as machine learning), the anticipation of the creation of new algorithms through the AI learning is usually known to the producer at the time of original sale. More so, this knowledge of its ability to learn is usually reflected in the marketing strategy and pricing of the product. 

For instance, a producer may assert to a buyer that its AI camera which is originally designed to identify a harm will subsequently be able to not only identify but alert him of such harm. Thus, even though he (the producer) at the time of the sale would not know specifically how the AI algorithm might evolve in the future, the fact that he knew it will evolve will make him liable if it turns out that the latter feature produced by the AI algorithm does not evolve or is harmful. In the above scenario, the producer will be liable for misrepresentation which was not caused by his own algorithm. 

3. Intellectual Property: Similar to the question above, is an AI capable of having intellectual property rights over its creation? Put differently, should intellectual property rights be attributed to an AI over its creation? Practically, who owns the IPR over the Non-Fungible Token (NFT)-based digital painting produced in part by a humanoid robot named Sophia? Does the machine artist Ai-Da have a claim on its drawings and sculptures, or do the rights belong to the person who developed her? 

There have been recent investigations on the above subject. While some writers have favored the attribution of IPRs on AI over its creation, it is important to note that IP laws only attribute IPRs to humans and where the law is silent as in the Australian Patent Law, the Court has restricted its application to only human creations. Moreover, AIs are neither natural nor legal persons. In a popular case, the Australian Federal Court had ruled in 2021 that an AI can be considered an inventor, its producer remains the patent owner. On Appeal, its decision was overturned, finding that an inventor must be a natural person. The same position is held in the UK, Germany, US and will be favored by most common law jurisdictions.

Conclusion 

The rapid deployment of AI is testament of the evolution of computer science. Law on the other hand, though sluggish, have emerged to regulate AI usage. It may be safe to conclude that Law is not proactive, but rather responds to changes. But in between these responses, lies the relationship between it and the growing influence of AI.