A wide range of practical and useful applications has been made possible by the rapid development of artificial intelligence (AI) and machine learning (ML) in recent years. Technology has indeed made life a better place. This review attempts to raise awareness about some potential dangers that the world might face if artificial intelligence is fully developed and also suggest possible ways to enjoy the important benefits of AI.
The article, "Artificial Intelligence and how it can be regulated" is herein referred to as "the article being reviewed".
The main purpose of the article being reviewed as spelled out in its introduction is a proposition that AI Governance/Ethics ought to be the first point of call before regulating the system.
As stated in the article being reviewed "To rush into an outright ban or overregulation of AI would result in not only a delay in the introduction of innovative ideas but can discourage inventiveness and innovation and an undue restriction to societal advancement. Eric Schmidt has described this to be the chilling effect."
I took some time to research Eric Schmidt's contributions to AI and I found a recent interview where he shared possible innovations to expect in 2027. Eric shared the breakthroughs AI would bring to Biology, predictions, automatic decision-making, and more. However, he also explained "the human element" in AI, which he said would be the solution to the challenges posed by AI.
The human element simply refers to the collaboration of humans and AI to produce more efficient results. In my opinion, this collaboration is only possible when AI is not fully developed. In Eric's words, there are so many things about AI which we as humans would never fully understand due to the accelerated thinking of an AI and its access to a large database of content. Instead of a collaborative relationship, it would be a psycho-conditioned AI-dominant relationship.
As opposed to the article being reviewed, this review proposes that AI Governance/Ethics would be impossible without strict restrictions on the fast-developing innovations of AI. The dangers of AI are more frightening in its developed state than in its primitive state.
The Nature of Technology (Smartphones as a Case study)
Over the years, we see how technological innovations evolve from being merely useful tools to our everyday companions, even becoming an essential part of our existence. Let's take tech innovations on solving the problem of communication as a case study.
According to the Forbes report, the first innovation of tech for communication in the 19th century is the telegraph machine. In 1849 an advanced innovation was designed called the Telephone, which within 50 years of use because an essential gadget in every office and home. The first handheld mobile phone was introduced in 1979 by Motorola to make communication easier and more efficient. In 2007 the tech industry experience a breakthrough in communication with the introduction of the very first Smartphone by Steve Jobs of Apple Inc. This was a huge turnaround, not really for the tech industry but for humanity. Smartphones dug deep into our existence and they became the next closest to us after our clothes.
Smartphones went beyond communication to a tool used to extract so much more personal information from us than we can imagine, they tampered with our data privacy, they are GPS trackers on us, they control the movement of our money and so many other important parts of our everyday lives. Due to the computing functions fused with the telephone functions to make a smartphone, they have also become mobile shops or offices where many carry on their business including all financial services.
The case study above shows how advancements in technology changed from being merely useful tools for communication to being a necessary component of our existence.
What would be the fate of the world if essential decisions of our lives are informed and executed by machines who through the study of patterns, trends, and human behaviour stored in a data chip have become intelligent enough to self-serve?
What is AI in simple terms?
Artificial intelligence is the imitation of human intelligence by machines. Indeed it has been a great contribution to modern civilization and this review is not aimed at downplaying its usefulness, instead, it is aimed to point out some risks which we are open to if AI allows running a frolic of its own. AI does not only seek to imitate human behaviour it also seeks to have a much more accelerated thinking than the human brain.
AI vs the human brain
"Oversight may be achieved through the introduction of mechanisms such as a human-in-the-loop (HITL), Human-on-the-loop (HOTL), or Human-in-command (HIC). While HITL (which implies the direct intervention of humans in every decision of the system) is almost impossible and is not desirable, HOTL (which refers to human intervention in the design cycle) and HIC (which is the capacity of humans to oversee the entire activity of the AI and to determine when and how to use the system) are most preferable."
The above excerpt from the article being reviewed shows a possible way to keep AI under the control of humans. I think this is only possible when AI is not fully developed. One pivotal feature of AI is "Automatic Decision Making". The development of AI is on the basis that humans have so many limitations which can be bridged by AI. The development of AI is not limited to innovations conceived by humans, it also covers innovations by AI itself. Innovations by humans could have a HIC mechanism for control. Would it be possible for humans to predict possible subsequent innovations by AI and control them? HITL, HOTL, and HIC are limiting the full advancement of AI, in the long run, they would be expunged for AI to take its full cause.
To keep humans in control strict regulations must be put in place to restrict the innovation and development of AI. This way we can enjoy the benefits of AI and also correct its excesses.
It is commonly argued that the human brain interacts with its environment and other human behaviours it is exposed to, to determine what it considers right or wrong which is the basis of a good rational decision. This process is also what AI goes through to determine its decisions.
I believe that there is something deeper than the human brain which may be impossible for an AI to simulate, the conscience. The conscience helps us balance what our brain says with what our heart says as a human thereby producing much better decisions than merely rational decisions. Hence, the need for strong restrictions on the advancement of AI.
Although the conscience may appear abstract, it is recognized as Natural Justice, equity, and good conscience in Nigerian legal jurisprudence. Humans take this into account when making decisions that an AI may find difficult.
Something to think about
There are times when we, as humans with both our brains and conscience, find it difficult to stop doing something wrong. How much more difficult would it be for an Artificial Intelligence without a conscience who has found purpose in doing evil?
AI vs Regulations
However, to prevent premature regulations of AI, we advocate for policy formulations that reflect these ethical standards discussed below. The European Commission published The Ethics Guidelines in April 2019 which provide that AI technologies must be Trustworthy. The article being reviewed went further to break down the 3 elements that make up a trustworthy technology which include; Lawful, Ethical, and Robust.
Our limitations as humans to fully understanding the operations of AI in its developed state stands as a threat to whether or not AI technologies can be trustworthy.
Technology has also generally proven to be the most stubborn of all industries to regulations and regulatory organizations. As technology evolves so do hackers and cybercriminals. Considering the power of secured VPNs, anonymity and the dark web it is difficult to keep technical innovations under regulations.
Although intelligent, artificial intelligence is exercised by machines and software that are vulnerable to cyber attacks, hacking, and other malware that can be difficult to control. This is why it is necessary to regulate the evolution of AI.
AI vs Terrorism and Conflict
Before I even start, imagine a world where we have several weaponized nuclear warheads controlled by Artificial Intelligence. I'm almost certain Nigeria can not afford to have one. It could be the beginning of the end of the human race. Technological innovation for protection could jolly well be another way to launch a more deadly attack.
In the interview with Eric Schmidt's, referred to in the introduction, reference was made to Nuclear warheads with automatic decision-making. This could put the world in grave danger.
Conclusion
Prof Stephen Hawking, one of Britain's most eminent scientists, stated that efforts to create thinking machines pose a threat to our very existence. In an interview with BBC, he mentioned that primitive forms of artificial intelligence have proven very useful. Once humans develop Artificial Intelligence, it will take off on its own and redefine itself at an ever-increasing rate. Humans, who are limited by slow biological evolution, couldn't compete and would be superseded.
If we would have surprises in the nearest future, let's make them pleasant and not the likelihood of an AI robot, in the face of a forthcoming election, saying 'AWALOKAN'.
My big ups to Faith Jane Adekunle for reviewing this paper. Thanks to Joel Odili ACIArb (UK), Joel Adeyemi Adefidipe, and Oreoluwa Onafowokan for the inspiration to write this review, Check out their work on Artificial Intelligence and how it can be regulated.
Comments