Since the invention of the computer, people have been fascinated by the idea that one day it can become smarter than humans. A conversation published on Internet several weeks ago has made people believe that this future may be closer than we thought. The transcript of the conversation between a human engineer and LaMDA, the artificial intelligence that as it seems, has its own thoughts, even feelings, is the last case that stirred the waters regarding the development of artificial intelligence (AI) and its real advancements.
Google’s engineer who published the transcript on the Internet was sent on vacation, and the company said he is under suspension because he has published confidential information. Google also completely denied that LaMDA is aware of its existence and has developed feelings.
Through social media platforms, mobile applications, autonomous vehicles, and smart home devices, AI has been present for years and is making people’s lives worldwide easier. But, even though almost every one of us is carrying AI in his pocket, installed in mobile devices, and is asking for its assistance every day – answers to questions and directions to locations – very few people actually understand what AI represents, the ways it functions and whether the fear that the AI might become smarter than the one who is using and controlling it is justified.
Meta.mk spoke with Dr Hristijan Gjoreski about AI development and where this system is applied in the country and abroad. Gjoreski is an Associate Professor at the Faculty of Electrical Engineering and Information Technologies (FEIT) at the Saints Cyrill and Methodius University in Skopje and a Chief Data Scientist at Emteq Labs, England.
His interests are AI and machine learning applied to the creation of intelligent systems and smart devices that are carried – smart glasses and smart bracelets. He participated in more than 15 national and international projects, and currently, he is a manager of the European project „WideHealth“. He is an author of more than 80 scientific publications, out of which 19 were published in international magazines with an impact factor.
What is artificial intelligence (AI)?
If we start with the name itself, “Artificial Intelligence” (AI) would denote any intelligence that is not natural. Next, the term natural intelligence should be defined and that is not simple at all. This is why my viewpoint and explanation of AI are biased toward the engineer’s definition i.e. the creation of intelligent systems that will assist us in our everyday lives, thus making them easier. In other words, AI is a modern tool of the digital era, that will be used to move society, progress and civilization forward, through the development of smart and intelligent computer systems. Hence, AI is a modern scientific discipline which appears and develops with the development of electronics and computer sciences. I’m certain that if you ask a philosopher, psychologist or cognitive scientist, each will provide a different viewpoint and definition of the term artificial intelligence.
The term AI was introduced in 1956 by Professor John McCarthy with his US colleagues. Since then, AI as science is followed by many controversies and is a mystery only because of its name.
There is a well-known adage by Professor John McCarthy that „As soon as it works, no one calls it AI anymore…“ In some way, we want AI to be something that we will strive towards, but we’ll never be able to reach it. Actually, everywhere we are currently using AI, we are using different phrases, so instead of AI we use words such as “smart,” “intelligent,” or “autonomous.”
One cannot talk about AI without mentioning Alan Turing who is considered to be the father of computer science and, indirectly, AI. The first idea about machines that will be calculating and thinking came from Alan Turing. He has also defined the well-known Turing’s AI test that will test whether a human is able to recognize if he is communicating with a human or a machine (i.e. AI). As he says, if we arrive at a phase when a human will not be able to recognize whether he is communicating with a human or AI at the other end, then we reached AI.
People often have a twisted view when it comes to AI, most likely due to the films that deal with this subject i.e. the first association is robots that look, talk, think and operate like humans. But, at the moment, AI and the people that are working in this area are not striving to achieve this at all, and in the near future, i.e. with the current technology and algorithms – it is not possible.
Therefore, we should stop thinking about robots when it comes to AI. Instead, the first association should be software i.e. smart algorithms, smart devices, and smart systems.
How developed is AI today?
Currently, we are in a phase where AI is actually machine intelligence i.e. computer intelligence, by computer and sensory systems. It represents the intelligence of the data gathered by these systems. This type of AI, which is creating algorithms and systems that function only for a certain type of task, is called Artificial Narrow Intelligence. At the moment, we are in the era of this type of AI i.e. AI that plays chess, AI that is able to predict problems with machines before they happen, AI that knows how to recognize our activities and to measure calories, AI that knows how to drive a vehicle, etc.
The next phase would be Artificial General Intelligence, which is defined as intelligence as smart as the people. This means the absence of AI and algorithms for every task separately, but general algorithms that would adapt to various tasks, for example, AI that knows how to drive a car and knows how to play chess.
The last phase would be Artificial Super Intelligence which can be defined as AI that is much smarter than people.
If you ask how it would look like if an AI drove a car, most people would envision a robot driving a car, but actually, the very car would be AI. Even these days we have commercial (not just prototypes) cars that can drive themselves more than 90% of the time (mostly on a highway and a well-marked road). Actually, it is an excellent example of AI put into practice – it’s the sensors, data, and algorithms that are steering the car.
Where is AI being used, and where is it applied in our everyday lives and people’s activities?
Today’s AI algorithms (machine intelligence) are all around us, especially in the digital world, in each use of mobile telephones, computers, cars, smartwatches, and smart home devices. One of the first successful AI algorithms without which we will not be able to function normally today is the email filter for spam messages. Most of us aren’t aware that over 85% of all messages are spam which isn’t reaching our mailboxes thanks to AI algorithms.
Today we are witnessing advanced AI applications in our everyday lives, such as digital personal assistants (Apple Siri, Google Assistant) that with the use of smart AI algorithms are able to recognize what we are saying or asking and then perform a certain task (to schedule a meeting, to create an event in our calendar, to send a message) and, in the end, to reply using a human voice. These smart assistants are already capable of “hearing and understanding” one language and responding in another, which allows communication between people that speak different languages.
Probably the most advanced AI we have and is being developed today is the AI that can autonomously steer cars. We are already seeing commercial vehicles (Tesla) that are capable of driving on their own (under certain conditions) with the assistance of AI and an enormous number of sensors and cameras. There are prototypes of cars with AI that are completely controlling them and are showing excellent results, but they are still being developed. Probably in a short while we will have to change the laws and rules of driving to adapt to autonomous vehicles.
Is someone deciding and controlling how AI will develop?
Like any scientific discipline, the direction the development will take is defined by people working with AI i.e. scientists, researchers, engineers, and enthusiasts. Currently, AI is in a phase where it can damage the people and the society we live in. Even though we are talking about computer systems, algorithms, software, and data, still the decisions that AI makes can have powerful consequences that we have to be wary of (e.g., a traffic accident caused by an autonomous vehicle – AI). This is why lately we are seeing more and more debates and discussions about introducing rules and ethical norms on AI.
An example can be provided with autonomous vehicles which are driven by AI. There is a debate that is calling for legal solutions and rules about who will take the blame if there is a traffic accident with an autonomous vehicle steered by AI. There are a lot of questions raised, for example., what if the sensors or the cameras have malfunctioned, what if the algorithm that is making the decisions was poorly programmed, what if the other vehicle that was driven by a human has caused the accident, etc.
Another example that can be provided is the armed autonomous flying drones that would be used for military aims. At the moment, we have the technology for one such flying drone equipped with weapons, to recognize its target (a human) and eliminate it. The only thing that is preventing the development of this technology and its use is the ethics and the rules we impose as humans.
Actually, several years ago, there was a public protest by almost all world scientists working with AI (through debates, letters, etc.) to stop and not finance research of this type in order to avoid similar misuses in the future.
This case resembles a lot of the nuclear research before WWII when the main goal was researching efficient and cheap sources of energy, but we all know how this research was redirected toward producing nuclear bombs. Therefore, we have to be careful with the AI and the intelligent systems we are creating and start thinking about the consequences if there is a mistake or a misuse.
What is the degree of AI development in Macedonia?
AI in the country is mostly represented in the scientific and research centers, such as the universities. I can confirm with certainty that several faculties at UKIM are working on AI and among them is the Faculty of Electrical Engineering & Information Technologies (FEIT) where I work as an associate professor. At the moment, FEIT is steering the European Twinning Project – „WideHealth“, where we are working with foreign partners on AI’s use in the electronic health sector.
Apart from the scientific and research centers, lately, there are more and more IT companies that are creating small teams that are working on applying AI algorithms to data (Data Science).
Specifically, I and my team are working with an English company named Emteque, and we are developing AI that is implemented in smart glasses, with the help of sensors installed in the glasses to recognize the activities, facial expressions, and emotions of people who wear them, therefore providing help in the everyday activities and needs. At the moment we have a prototype of glasses that are used for collecting data for FEIT, we are analyzing the data and are creating these AI algorithms, which very soon will be applied in our everyday activities.
Also, with the help of the Fund for Innovations and Technological Development (FITD), a significant number of small startups that are working on AI utilization in the industry are being supported.
FITD has also started with an initiative for the creation of a national AI strategy, where various professors, engineers, industry representatives and the NGO sector have taken part.
Republished from: Мета.мк