Perhaps in the nearest future, it seems clear that Artificial Intelligence (as AI after now) technology will be considered as a property, as it is in some advance countries today, where 1 Robot perform many activities on it’s own based on the general commad or training it takes. Yet the technology of the day despite it’s undisputable growth, no AI so far could be said to be free of all unintended consequences, it may lose control sometimes, or developed some technical issues or been hacked or attacked by virus. In all these cases unlikely things may happen to the case a law firm handles, or the car I robot was driving or self driving car to cause some accidents or while reviewing contractual agreements the AI skipped some important information to be dealt with may turn very unfortunate to the client, or due to it’s incapablity of self determine and control be used to commit an act which is against the law of the land whether intentionally or mistakely, at this point the question of who will be held liable arises. these and other related things are the salient legal and ethical issues that this article is set to touch with highest economy of words and make to some recommendations which can be used to address the predictive issues.
The artificial intelligence in 21st century
as seen in the previous article the AI of 21st century has came with remarkable innovations today we have robotic cars which don’t require a driver to control or supervise them, smart machines that process a large amount of data that a human being can’t be in position to perform, the machines can actually think, although they can never have a sense of humor, fall in love, learn from experience, know how to distinguish the good from the bad and other attitudes of the human.
AI today is the combination of various technologies that give chance to robotics to understand, learn, perceive or complete human activities on their own, it’s tool is majorly concerned with emphasizing robotics which portrays human behaviors. But however, Articial Intelligence may fail out at some points due to differences in human brain and computers and sometimes it may developed certain technical issues that may affect it’s performance.
One fundermental thing to observe is AI of this 21st century is still not that absolute autonomous as some writers believe that Till now(21st C), AI machines are not able to control their process, they still require the intelligence and mind of human beings.
LEGAL AND ETHICAL ISSUES
Today’s technology give it chance to many technological instruments to be used in delivering massages and perform some certain contractual duty, an I robot can be used to deliver or receive some contractual agreements, notices, letters etc. It can listen to the client and prepare contractual decuments by itself,other AI machines like eBrevia, Leverton can review and amend contract or highlight the area of ambiguity or inconsistency in the contractual decuments.
Assuming, the validity of contract’s offer to a wrong party, or the contractual agreement that can’t be fullfil at the circumstance due to defect with the necessary technological instrument, for example A is under contractual obligation to publish B’s article on his Facebook page, unfortunately A’s account hacked by an unknown hacker just some minutes before the publishing time.
The Basis for the law of agency is summed up in the Latin maxim ‘Qui facit per alium, facit per se’ that ‘He who acts through another, acts as himself’. An autonomous AI agent possesses the ability to operate without the direct intervention of human principal or other human agent, while the non autonomous is always under the control and supervision of it’s operators, autonomous agent determines terms of the contract (such as price, taxes, and charges, or terms and conditions) independently of the operator.
What is the legal status of a contract entered by an I robot agent (as machine that can take a decision) beyond it’s given power, or take a decision that as an agent has no power to. And when the robot because of it’s inability to distinguish right from wrong did something that breached the contractual agreement in the course of performing some part of the contractual task?
In all the stated events, who will be hold liable believing that the harm caused may have been unforeseeable from the perspective of a specific component manufacturer. And could be a defect from the communication line of the robot with it’s operator, or any other unforeseen defect, to be fair, will the other party only suffer the Lost of the contractual instrument? in what circumstance the contract law doctrines of ‘non est factum’ and ‘unilateral mistake’ could avail the principal human.
Our task here is not to deal with the theoretical issues as regards to the contract law and agency whether or not the AI is use as an agent (autonomous agent) or a mere contractual instrument. Hence we should not labour ourselves discussing the theoretical part of the arising issues, an interested reader should read Samir Chopra and Laurence F. White’s ‘A Legal Theory for Autonomous Artificial Agents’ the learned writers have done justice to those theoretical aspects.
Criminal liability of some offences in our jurisdiction require both mens rea and actus rea(Strict liability offences). But before we get there which is very unlikely to get the two elements of offence in case of AI robot. Let us first appreciate whether AI robot can done something which Criminally wrong under state or federal laws, like stealing, extortion, robbery, sexual assault and their likes? And whether AI robot can be Criminally held liable under Nigerian criminal justice system?
The AI technology is designs to learn and accept instruction without the ability to discern what is right or legal with wrong or illegal. It has no different with a gun ammunition that if directed could turn and shot it’s owner, it’s ability to think and make a decision is said to have limitations as a writer Harry Surden observed in his article ‘Artificial Intelligence and Law: An Overview’;
“current (21st century), AI technology has its limitations. Notably, it is not very good at dealing with abstractions, understanding meaning, transferring knowledge from one activity to another, and handling completely unstructured or open-ended tasks”
This will qualify AI robot to be free from criminal liability. Since it can’t commit an offence on it’s own but under the direction of a criminal human being or when something goes wrong in it’s control, manufactures or communication.
Another important reason of this writer is our criminal justice is designed not only for the purpose of prosecuting and convicting a defendant but also to deter the commission of the same offence in future by the same defendant and his contemporary. AI Robot lack the ability to learn from experience of the fast and it equally incapable to control itself from committing an offence of which it’s controller commands.
Turning back to the mens rea and actus rea as elements of offence, in the humble opinion of this writer, if anything, as the circumstances should clearify more, it is the operator of the AI that has the two elements of offence, by controlling and directing or programing the AI to commit the offence in question. Because a person who controls and drives his car cannot reasonable hold it as defense that he is only a driver who intend to drive recklessly but the real killer is the car he was driving. But what if the operator/manufacturer did not program the AI device to commit the offense, and did not order it to do so?
Very noteworthy, Gabriel Hallevy in his book ‘When robots kill artificial intelligence under criminal law’ (one of the very interesting book ever written on Artificial Intelligence Law and Criminal liability) holds otherwise as he imaginarily predicted the circumstances that an AI robot can be held criminally liable for offence of manslaughter and be convicted and sentenced accordingly. He called it ‘machina sapiens criminalis’
Putting it in his words, the learned writer states;
“After the court has convicted the ai robot for manslaughter, the criminal process continues as if the offender were human, and the next stage is that of sentencing. The prosecution and the defense may argue for a harsh or lenient penalty, and eventually the court decides the appropriate penalty based on the relevant circumstances of the case….After its imposition, the penalty must be carried out. As previously noted, the imposition of imprisonment on ai systems takes the form of depriving them of their liberty to act by restricting their activity for the required term and under strict supervision. During this time, the ai system can be repaired to prevent the commission of further offenses”.
To him technological capabilities of the AI are what needed to meet the requirements of it’s non strict liability offenses, what forms a machina sapiens criminalis.
Apropos to our submission on Criminal liability it could be sensed that AI Robot can also be used to commit tortious acts like assult, battery, tresspass to person, tresspass to chattel, convertion etc. It may also cause economic damages which is generally recoverable in torts law And can also be mistakely as self driving car hit someone or tresspassed to someone’s property.
Historically, tortious liability was meant to protect human, his body and property against the intrusion of another human, that gives a man numbers of causes of action to protect his legal territory against any intruder, it also protects one to be harmed by someone’s breach of duty of care. The theories of torts are flexible enough to handle artificial agents include existing doctrines relating to liability for wild and domestic animals, children, unpredictable actors under supervision such as prisoners and even slaves, and ultrahazardous activities.
The question of whether AI device is civilly liable for the act done by it under our tort law is as well very unlikely to hold the device liable, because ‘product liability and negligence’ will be the primary causes of action rather than the acts of the device. Because a manufacturer owes a duty of care for all their users, not to produce injurious factors and if their is anything without proper guide for it safety use, also a person knowingly using or dealing with dangerous instrumentalities or machines is required to exercise duty of care such as guarding, covering, or protecting them commensurate with the risk or danger of the injury involved; failure to do so may constitute actionable negligence for injuries caused as a natural and probable result.
Although, the AI devices possessed the capability to act without direct operator, still the operator, manufacture or the owner as the circumstances may be will held liable on Strict liability based on the rule in the Rylands v.Fletcher and perhaps under the principle of ‘Respondent superior’.
CHANGING THE LAW; to address the unaddressed issues.
Law in our modern setting has initial reaction when faced with novel innovations to curtail the predictive vices and to appreciate the best part of the innovations through legislations and judicial interpretations, AI divice has to destine this fate in our jurisdiction. As many countries have already enacted legislations to regulate the use of AI in certain fields.
Law professors, I. Giuffrida, F. Lederer, and N. Vermeys in their very riveting work titled; ‘A Legal Perspective on the Trials and Tribulations of AI: How Artificial Intelligence, the Internet of Things, Smart Contracts, and Other Technologies Will Affect the Law’ had this to say:
“AI age starts with traditional legal concepts increasingly applied to new and previously unforeseen circumstances impelling legal change. This has happened before, of course, but the AI age will not only be immense in scope it will also proceed incredibly quickly. Our legal systems tend to be reactive and not proactive, especially when we cannot predict what the future will be like”.
But this humble writer believes that our legal system has to be both reactive and proactive, it has to address the current issues at hand and the likely predictive issues in the near future. We shouldn’t live like blinds who could only believe the arising of sun when it struke our faces. Let use the past and present experience to address the future or we will be doom to repeat the mistake of the past.
To borrow the famous quote attributed to Spanish philosopher George Santayana, “Those who cannot remember the past are condemned to repeat it”
Only if one could relate, is the cybercrime Act that came in 2015 addresses the solid issue of cybercrimes in Nigeria, can we say the act came in on time? Can it be said that the act is preventive or proactive? As it’s crimes strategies is improving with the development of technology. If so does the law provides adequate mechanisms to track and apprehend the cybercriminals?
We can’t make laws in vacuum and stick to wait for the unknown to be known as we have many unfulfilled holes in our legislations but the reality of AI technology is the reality of the day, which we shouldn’t risk to wake up late as we did for many issues in the past.
In September 1, 2017 New-York Times Op-Ed, Oren Etzioni proposed three rules that he believes should apply to AI
I. an AI system must be subject to the full gamut of laws that apply to its human operator;
II. an AI system must clearly disclose that it is not human;and
III. an AI system cannot retain or disclose confidential information without explicit approval from the source of that information.
For the first recommendation, the AI device shall be subject to the law but not to be held liable aside it’s operator,user or manufacturer human.
Adopting the Oren Etzioni’s recommendations we believe the following things has to be made in time;
I. We should have an act that will regulate the use of AI Technology, it’s marketing, and most importantly compulsory registration of all AI private device.
II. If anything AI should only be use as contract agent like The then slave agent who was merely as intermediaries but not to owe any contractual obligation.
III. Although in not distant future (in our jurisdiction too) AI devices will be treated as autonomous agents who will have the capacity to enter contract and to determine contractual terms on their own right. The were not to be recognised as legal persons. They should therefore lack the capacity to sue and to be sued in their own name.
IV. Unmanned AI robot and first misbehaved robot should be subjected to supervision of authorities similarly to a post-convict under section 3 of the Prevention of Crimes Act.
V. The licence of an AI operator and manufacturer or importer as the case may be should be subject to review from time to time.
VI. To avoid criminally use of AI, it’s role and usage to the operators should be forced to be transparent, justifiably without prejudice to their right of privacy.
VII. The AI device should also have the protection of the law from the evil acts, like destruction and stealing, to mention few.
Inevitably, every technological advance is accompanied by legal and ethical questions. Because we must all have the regulations, and the way of doing things. As the legal and ethnical issues surrounding AI technology is likely if not definitely to crop up in very near future. Addressing such issues is unavoidable as equally as the acceptance of the technology itself.
One may say why the AI technology now when we still have unresolved problems of insecurity, insurgency, corruption and many more? and or how many Nigerians can afford the device?
Yet the AI technology could fairly be used in fighting the insecurity and corruption. And we have to look at it predictive impact on the society and not it affordability, since we can’t say that no one can afford it in Nigeria.
To avoid such unwelcoming events, we must proactively prepare for them, and not to wait until we have understood all there is to know about these technologies because their improvement is increasing every day, that’s why while enacting any law concerning technology and it’s used we have to be very conscious so the technology will be a blessing rather than a curse.
ABOUT THE AUTHOR
ALKASIM ABUBAKAR Is a law student of AHMADU BELLO UNIVERSITY Zaria. He is a paralegal, political analyst and critic, an educationist, a season and inspirational poet, an avid reader and a prolific writer, he has written numerous articles on various areas of law and contemporary issues and he is widely published. He can be reached via; [email protected]
For knowledge and Justice