The learning algorithms used by social media sites are becoming more and more adept at recognizing objects, people, and even the detailed characteristics of those objects and people. Given the speed at which artificial intelligence develops, it is possible that we will soon encounter surprising innovations.
Every time you browse the internet to read news about artificial intelligence today, you can learn about new artificial intelligence systems that manage to do many things more successfully than humans. Artificial intelligence can better diagnose cancer types than human doctors, create advanced AI algorithms and beat world champions in games like chess. Examples like this can lead us to believe that artificial intelligence is not what it can do better than us. The evolving features of artificial intelligence lead to hope in many areas, as well as fear. While many believe that the development of artificial general intelligence can greatly benefit humanity by raising our standards, some believe that development can lead to global disasters.
While the superior development of artificial intelligence remains a topic for discussion, before we worry about the future, we need to understand how the current artificial intelligence is making progress and how far real development is from today.
What are the latest developments in artificial intelligence?
The Internet is full of stories of stunning applications born out of today’s artificial intelligence research. There are many examples that resemble artificial intelligence that can diagnose cancer, mimicking human consciousness and logic. For example, the learning algorithms used by social media sites are becoming more and more adept at recognizing objects, people, and even the detailed characteristics of these objects and people. Modern computer technology can identify a person’s posture, facial expression, and jewelry in any image. This gives AI systems the ability to detect images the way humans do. AI systems can go beyond identifying people from images, even analyzing fine details to discern qualities that are not clearly visible.
Another example of artificial intelligence that exhibits human-like characteristics is the ability of artificial intelligence to understand speech, i.e. NLP (perception of natural language). AI specializes in understanding text and conversations, which are part of apps such as chat and virtual assistants on smartphones (Siri, Cortana, etc.). With these developments, the difference between human intelligence and artificial intelligence is rapidly decreasing. This may give you the impression that advanced AI systems may not be much in the future. However, it is important to understand that qualifying as” advanced artificial intelligence ” requires more than just performing certain tasks better than humans.
What exactly is artificial general intelligence?
Simply put, artificial general intelligence (AGI) can be defined as the ability of a machine to perform any task that a person can do. Although the above-mentioned applications emphasize that artificial intelligence is more successful than humans in some jobs, these systems are generally not intelligent, they are exceedingly good only in a single task assigned to them. So while an AI system is as effective as a hundred trained people at performing a task, it can be extremely challenging when asked to care for a five-year-old.
For example, although computer vision systems are skilled in the field of visual information, they cannot convert or apply this ability to other tasks. In contrast, people can be successful at performing more tasks than any of today’s AI applications, although they are sometimes less competent at performing these functions.
Artificial intelligence needs to be trained to gain experience, while people can gain experience in a much shorter time. In addition, a system with artificial general intelligence will have a minimal learning process and will be able to apply the knowledge it has learned to other areas. For example, my AGI site, which has only learned one language, can learn languages with common roots and similar syntax using NLP. Such an ability will make the learning process of intelligent systems similar to humans, allowing the machine to gain multiple areas of expertise, while significantly reducing training time.
Can artificial intelligence gain the ability to think?
Artificial intelligence systems, especially artificial general intelligence systems, are designed with reference to the human brain. Because we don’t have complete knowledge of our brain and its functioning, it’s very difficult to imitate the way it works. However, it is theoretically possible to create algorithms that can replicate the complex computational capabilities of the human brain. As proposed by the Church-Turing thesis (simply put), having infinite time and memory can solve all kinds of problems algorithmically.
When will it be possible to achieve artificial general intelligence?
Even if copying the workings of a human brain is theoretically possible, it is not feasible for the time being. That’s why we’re far from getting this kind of artificial intelligence. But given the speed at which artificial intelligence develops, we are likely to encounter surprising innovations soon. While some experts say artificial intelligence will develop as recently as 2030, they think artificial general intelligence can be achieved by 2060.