Artificial Intelligence (AI) is a term that would remind us of the greatness
of robots or systems which are often synonymous with the ability to act like humans. With the extremely massive and progressive development of technology and computer systems, instead of aiding man’s roles, artificial intelligence has been gradually used to replace human roles in certain jobs.
Although the development of AI has only been noticeably massive since a few years ago, history has proven that artificial intelligence has become the scope
of experts’ research long before the internet developed and touched every aspect of our lives as we know it.
H.A. Simon (1987) defined artificial intelligence into the areas of research, application and instructions associated with computer programming to do things which, in man’s view, are deemed intelligent. Furthermore, rich and knight (1991) defined artificial intelligence as a study on how to make computers do things which are currently done better by humans.
According to BBC, put simply, AI is a “machine” capable of doing various things where, if done by humans, intelligence is considered required, such as understanding human language naturally, facial recognition in photographs, driving vehicles or even guessing what books we like based on the books we have read before.
Google Assistant which can be found in Pixel smartphones, or Siri in the hardware ecosystem of Apple, and Cortana in Windows operating system, are perhaps representatives of such technology. If we look even closer, the simplest AI can actually be found in calculators, or when we process some data using Microsoft Excel.1
Artificial intelligence system becomes even increasingly developed in
our economy and society, and is designed with increasingly improved capabilities to operate independently from humans’ direct supervision. The algorithm system in stock market dealings and autonomous vehicles with AI “drivers” which have been road- tested are all serious examples of how AI has had brushes with the law. AI which might become very close to interacting with humans are bots, computer programs on social media. AI bots have become increasingly advanced and sophisticated as they could interact in dialogs with real humans. This has garnered attention of cyber law experts in the US.
In 2016, in the US, an AI bot known as Jill Watson – tasked as an assistant teacher in an online course in Georgia Tech University – managed to fool and trick students into thinking that it was an actual human being.2
A more serious example is the widespread use of Pro-Trump political AI bots on social media several days before the US presidential election in 2016.3 Because of the incident, the US cyber law experts pushed the Congress to amend the law in order to restrict and regulate the use of AI.
Despite AI’s potentially extensive legal impact within the society, particularly concerning the legal accountability pertaining to it, the Indonesian legal system itself has not expressly regulated the issue. This has raised a number of questions, such as how the legal system can ensure that victims are properly compensated if an AI causes a physical or economic damage? In addition, there is also the question of how AI can have the same treatment as humans when it comes to legal accountability? Or to what extent should a system owner be held accountable for the actions of such autonomous AI system?
In order to address these questions, at least a few things need to
be considered. Discussing legal accountability means also discussing the capacity of a legal subject to be held accountable.
Explicitly, although AI is capable of taking actions included as legal actions, it cannot be identified as a legal subject. Therefore, an alternative is required while still associating the AI’s actions with the legal subject that is the AI’s owner.
We can find an alternative by conducting analogical analysis or interpretation. Analogical interpretation provides interpretation of the wording in a legal regulation through analogies in accordance with its legal principles. Therefore, an incident which actually cannot be included in the definition of a certain regulation can be otherwise considered included by the wording of such regulation.
By using analogical interpretation, the association between an AI and its owner can be illustrated in the following examples:
1. A relationship between a pet and its owner
As the first alternative, we can draw an analogy between the relationship of an AI system and its owner and the relationship between a pet and its owner or user, as referred to in Article 1368 of the Indonesian Civil Code (“KUHPer”):
“An owner of an animal, or an individual who uses one, as long as the animal is available for his use, shall be responsible for any damage caused
by the animal, whether the animal is under his supervision and in his care, or whether it is lost or has escaped.”
In civil law, if a pet causes damage, whether the pet is under its owner’s or user’s supervision, or whether the pet is lost or has escaped, such damage becomes the responsibility of the pet’s owner or user.
Similarly with AIs, due to its autonomous nature, AIs can be considered analogous to that of a pet under its owner’s or user’s supervision. hence, if an autonomous AI performs any action unforeseen by its user or owner, the incident can also be considered analogous to that of a “lost or escaped” pet. Therefore, if an AI commits any action which, by law, impairs or harms other people, the owner can be held accountable, under civil law, for the AI’s action.
2. Relationship between an employee and an employer
The relationship between an AI and its owner can also be considered analogous to that of an “employee” and an “employer”, as set forth
in Article 1367 paragraphs (1) and (3) of the Indonesian Civil Code (“kUHPerdata”), which states:
Article 1367 paragraph (1)
“An individual shall be responsible not only for the damage he has caused by his own act, but also for that which was caused by the acts of the individuals for whom he is responsible, or caused by matters which are under his supervision.”
Article 1367 paragraph (3)
“Employers and those who have been assigned to manage affairs of other individuals shall be responsible for the damage caused by their servants and subordinates in the course of duties assigned to them.”
Considering the aforesaid provisions of such articles, we can draw an anthropomorphic analogy by identifying the attribution of “worker” characteristics in an AI system. Besides, the use and application of AI in daily lives are aimed at doing works which could actually be done by humans. If AIs are considered analogous to “workers”, the legal accountability could be attributed to its owner that can also be considered analogous to the AI’s “employer”.
Technological advancement is inevitability. To anticipate it, therefore, the law should be constantlyupdated to cope with technological developments. Nevertheless, the use of AI systems must always comply with the entire set of laws applicable to their human operators.
No matter how autonomous AI systems become, we cannot let go of their reins and let them “run freely” without setting boundaries and regulations for such systems. Finally, while it is true that the Indonesian legal system has provided means to fill the legal void through analogy, the alternative analogies proposed above are hoped to provide a brief overview of how this area of cyber law will evolve in the future.
2. https://www.nytimes.com/2017/09/01/opinion/artificial-intelligence-regulations-rules.html 3. Ibid.