Artificial intelligence – these two words are very familiar to us today. At TechnoVision, we have repeatedly told you about it. There are many articles on this topic, but almost all of them were about AI systems that are “weak” – that is, created to perform specific tasks and cannot go beyond their specialization. But if you remember and read carefully, we have also occasionally briefly mentioned strong AI, also known as general artificial intelligence (AI/AGI). In this article, we will now take a closer look at this so-called superintelligence. Without dramatizing or frightening, we will try to tell you about one of the most important technological challenges of our time in a simple and understandable way.
What is AI and how does it differ from conventional weak AI?
Strong artificial intelligence is a conventional name for machine intelligence that would have general mental abilities at or above the level of a human being. Unlike weak AI, which is strictly limited to solving only one task at a time, strong AI has universal intelligence and can apply its capabilities to any problem, not just one specific one. In other words, a narrow AI today can play chess or recognize faces in a photo, but it will not be able to learn a new task beyond what it was specifically programmed to do. A strong AI, on the other hand, would be able to learn and think like a human without constant help from programmers, and transfer knowledge from one field to another.
For better understanding, let’s give examples. Virtual assistants like Siri or voice assistants in phones are narrow AI that recognize commands and respond to them within a predefined set of functions. Even modern models like ChatGPT, which can support a meaningful conversation, are still an example of weak AI, as they are limited to the task of conducting a textual dialog and cannot suddenly learn, say, how to drive a car or cook a meal. Instead, an AI would have the intelligence and self-learning capabilities of a human: it would be able to read books and acquire new knowledge, reason, draw conclusions, solve creative problems, understand different contexts, and even show some creativity in new situations.
Please note that strong AI has not been created yet and remains a theoretical concept. Scientists and engineers do not yet know exactly how to build a machine that thinks like a human being, because the human brain and consciousness are extremely complex. At the same time, progress in the field of AI in recent years has brought us closer to this goal than ever before.
Intelligence is higher than human
If strong AI is equated with human intelligence, superintelligence is a step further. This is the name given to intelligence that far exceeds the capabilities of the human mind in almost all areas. In other words, superintelligence is AI that is smarter than the smartest humans in every important field: science, creativity, wisdom, etc. It’s important to note that it’s not just about computing speed – computers already calculate faster than we do – but about the quality and breadth of thinking. A superintelligence could find solutions to problems that humans cannot solve, come up with innovations and ideas that we would never have thought of, and do so very quickly.
Superintelligence is still a purely theoretical idea, but if humanity manages to create a strong AI (at the human level), it is likely that technology will go further, and AI will surpass us intellectually in a short period of time. Some researchers suggest that superintelligence may even have traits of consciousness, its own aspirations or motivations.
Although this is a controversial issue now, superintelligence would mean for us the emergence of a subject smarter than humans. In science fiction, such an AI is sometimes imagined as something like a “conscious machine” or even an artificial brain capable of independent development.
What opportunities will a strong AI have?
A strong AI (and even more so a superintelligence) will have extremely broad capabilities, as it will be able to learn and apply knowledge in various fields. Here are some examples of what such a system could potentially do:
Solving scientific and technical problems. A strong AI could become the best scientist in any field, as it can instantly analyze huge amounts of data and come up with new hypotheses. For example, narrow AI systems are already achieving significant scientific results: the DeepMind AlphaFold program made a breakthrough by learning to predict the three-dimensional structure of proteins with an accuracy of up to the width of an atom. This was considered an extremely difficult task, and AlphaFold’s success is helping biologists find new drugs and understand diseases faster. Let’s imagine that a strong AI can not only solve one such problem, but simultaneously work on thousands of scientific questions, ranging from the development of cancer drugs to new energy sources. Such a system could make new scientific discoveries almost every day, accelerating progress many times over.
Simulations and forecasting. Human intelligence is limited in its ability to imagine and calculate complex systems, especially if there are a lot of factors. A strong AI, with powerful computing resources and smart algorithms, will be able to model reality in fine detail. This means that it is theoretically capable of conducting highly complex simulations, such as predicting the development of the climate on the planet, modeling the economy of an entire country, or the course of epidemics to find the best strategies to contain them. Even further, superintelligence can model social processes or even individual societies to see the consequences of certain decisions.
Automation of complex tasks. AI will be able to act as a kind of universal manager for many processes. It will take on routine and even creative tasks in various fields: from production and logistics management to analyzing legal documents or composing music. Unlike modern programs that operate only according to a predefined algorithm, a strong intelligence will be able to make non-standard decisions on its own. For example, it could manage an entire smart city, automatically optimizing energy consumption, traffic flows, responding to emergencies, etc.
Creativity. Interestingly, AI has already learned some forms of creativity: algorithms generate paintings, musical compositions, and poems. However, they do so based on patterns learned from human data. A strong AI could create truly innovative ideas and content by combining knowledge from different fields. Just imagine movies or video games whose scripts were created by artificial intelligence based on audience preferences, or new types of art that humans can hardly comprehend. In addition, superintelligence is capable of generating inventions: for example, offering new technologies, architectural designs, and engineering solutions that humanity has never known before.
Thus, the possibilities of strong AI will cover almost all areas of activity. This is both exciting and challenging: how will we use such opportunities and what will they lead to? But before we move on to the impact on specific areas of life, in the next part of our article, we will consider one interesting question: Will artificial intelligence be able to go so far in its capabilities that it will create entire artificial worlds that will be indistinguishable from the real world?
