In this blog post, we will examine whether strong AI and super AI can become reality, focusing on current technological developments and future possibilities.
As the Fourth Industrial Revolution progresses, the world is changing at an astonishingly rapid pace. In this article, we will focus on artificial intelligence, which has recently become a hot topic among the various technologies driving the Fourth Industrial Revolution. Artificial intelligence has been a topic of interest for a long time. A famous example is the movie “The Terminator.” The movie depicts a world dominated by machines with advanced artificial intelligence. When the movie was first released in 1984, many people were unfamiliar with artificial intelligence and viewed the plot as an entertaining fiction. However, last year, artificial intelligence defeated the world’s best Go player, demonstrating its astonishing rate of development, and many people now believe that it will not be long before artificial intelligence surpasses humans. For this reason, artificial intelligence is worthy of attention, and many people are focusing on its future. The history of artificial intelligence is actually not very long.
AI, which involves teaching computers to learn various human abilities through computer programs, was discussed by scientists in various fields in the early 1950s and became a field of study. However, it has only been a short time since significant progress began. It would not be an exaggeration to say that most of today’s AI technology was developed after 2010. In 2012, an experiment was conducted where several pictures were presented, and AI was used to give titles to those pictures. In the first attempt, the AI made many mistakes, such as recognizing a man kissing a woman as a piece of candy and labeling the image “Woman Holding Candy,” and only correctly named 7 out of 12 images. However, by 2013, the AI was able to assign appropriate titles to all images.
In 2014, an experiment was conducted to distinguish between Chinese, Japanese, and Korean people using artificial intelligence. When shown 20 faces, the AI achieved a 75% accuracy rate. This is a remarkable result when compared to the fact that Westerners could not distinguish between the three nationalities with even 50% accuracy. In 2016, AI defeated the world’s best human player in Go, a game with over 10^100 possible moves, stunning the world. Considering that 69% of AI experts predicted Lee Sedol’s victory, AI had made a leap forward beyond even the predictions of experts.
Since then, AI has begun to make its mark in a wide range of fields. In 2020, the AI model GPT-3 was released, featuring 175 billion parameters and achieving remarkable results in natural language processing. GPT-3 demonstrated exceptional abilities in writing human-like text, engaging in conversations, and translating various languages. In 2021, DeepMind’s AlphaFold made significant progress in solving the protein structure prediction problem, opening up the possibility of groundbreaking discoveries in biology and medicine.
In 2022, AI began to be widely applied in various industries, including autonomous vehicles, medical diagnosis, and financial analysis. In particular, Tesla and Google’s Waymo successfully developed and commercialized self-driving cars that operate on actual roads using artificial intelligence technology. In the medical field, artificial intelligence led medical innovation with outstanding achievements in cancer diagnosis, drug development, and genetic analysis.
In 2023, the release of the artificial intelligence model GPT-4 further accelerated the pace of development of artificial intelligence technology. GPT-4 has been trained on more data than previous versions and is capable of solving more complex problems. Discussions on ethical issues related to AI have also become more active, and research and policies are being developed to enhance the transparency and accountability of AI.
In 2024, AI became even more deeply integrated into everyday life. Generative AI technology has acquired multimodal capabilities that enable it to comprehensively understand and generate various forms of data, including text, images, and audio. In particular, OpenAI’s GPT-4o has attracted attention as a model capable of integrating text, images, and audio. This model enables real-time voice conversation and visual information recognition, opening up new possibilities in various fields such as education, counseling, and customer support.
Additionally, in 2024, global IT companies began competing to develop their own AI models, marking the official start of the ‘model competition era.’ Google announced the Gemini 1.5 series, enhancing its memory-based long-term dialogue capabilities, while Meta, Apple, and Samsung also launched services integrating AI functions into their respective ecosystems. In Korea, Naver and Kakao have consecutively introduced services based on large-scale language models, enhancing the country’s AI competitiveness.
As of 2025, AI has moved beyond being a simple tool to become a key component in high-level tasks such as creation, decision-making, and design. Its practical application is growing in fields requiring specialized expertise, such as programming, legal consulting, and marketing strategy development. Some companies and institutions are adopting new organizational structures that treat AI as core personnel. In the medical field, AI is beginning to be used for personalized diagnosis and treatment planning, and reports and papers written by AI have reached a level where they can be reviewed academically.
Amidst these developments, scientists felt the need to classify AI and divided it into weak AI, strong AI, and super AI according to its intelligence and characteristics. Weak AI refers to all current AI systems. It lacks self-awareness and the ability to learn on its own, performing repetitive tasks according to stored algorithms. Therefore, it possesses partial intelligence within the scope of the stored algorithms. Even AlphaGo, which is highly skilled, is limited to the game of Go and thus falls under weak AI. Significant research is still required for weak AI to advance to the next stage of strong AI. To implement strong AI with the characteristics of self-learning and self-awareness, it is necessary to understand the human brain, which has a very complex structure. While it is relatively easy to implement tasks that humans find difficult, such as multiplying 10-digit numbers or performing complex calculus, it is challenging to create AI that can perform simple movements or possess intuition.
Strong AI can be divided into two categories: artificial general intelligence and artificial superintelligence. Artificial general intelligence is often referred to as strong AI or complete AI. Unlike weak AI, which can only perform functions that have been programmed into it, strong AI can learn all intellectual activities that humans can perform on its own. In fact, it is expected to adapt to the environment faster than humans and perform tasks more efficiently. Artificial consciousness refers to AI that can feel emotions and possess creativity. Therefore, it has the characteristic of performing tasks it deems necessary without human commands. According to Nick Bostrom, a Swedish philosopher and Oxford professor specializing in AI, strong AI could be realized around 2040, approximately 25 years from now. A representative scientific technology used to achieve strong AI is the use of the nematode Caenorhabditis elegans. C. elegans is the only living organism whose entire neural network has been mapped by humans. Scientists used this information to create a robot with the same neural structure as the organism. The robot demonstrated behaviors similar to those of a living organism, such as avoiding walls and searching for food. This experiment, which achieved characteristics similar to those of a living organism using only neural structure without the aid of artificial intelligence, demonstrated the potential of strong artificial intelligence. This is because if we can understand more complex and advanced neuron structures, we can create those organisms.
Super AI is an artificial intelligence that is far more intelligent than the most outstanding humans in all fields, including society and science. Humans will reach a stage where they do not need to do anything, as artificial intelligence will advance science and technology on its own and open up a more developed world. Super artificial intelligence is also related to the technological singularity. The technological singularity refers to the emergence of artificial intelligence that is more intelligent than the combined intelligence of all humanity. Once artificial intelligence develops to this level, humans will no longer be able to understand the principles behind the technology used by artificial intelligence. No one knows what artificial intelligence will be capable of at that point. If humans can create Wi-Fi and establish the academic concepts of quantum mechanics using their brains, super AI, with an intelligence level approximately 170,000 times that of humans, could lead humans to immortality and explore other planets. It could be thought of as the descent of a god. Nick Bostrom suggests that super AI could be realized in about 45 years, by 2060.
The current era is often referred to as the “age of change.” Every year, the science and technology used in artificial intelligence are advancing at a rate of 200%, and artificial intelligence has already surpassed humanity in many fields. However, current artificial intelligence is limited to weak artificial intelligence, which cannot learn on its own and lacks self-awareness. We are now at the stage of developing strong AI by giving AI self-awareness and the ability to learn on its own. Strong AI will be able to continue to develop its own intelligence, reach superintelligence, and possess abilities far superior to those of humans. It will only take about 20 years for AI to reach that level of ability.
The pace of AI development is so rapid that even scientists cannot fully predict its impact on humanity. While it is desirable to advance AI rapidly to drive the swift development of civilization, it is crucial for scientists to take responsibility for humanity’s future and accurately understand how science and technology will affect humanity. In this era of rapid change, we must clearly understand the potential and limitations of AI and explore ways to utilize it appropriately.
Human creativity and emotions may remain areas that AI cannot surpass, but the development of AI will open up new possibilities that we cannot even imagine. As we envision a future where AI and humans coexist harmoniously, we must monitor the development of AI and prepare for it. The development of AI has already begun, and it is an unstoppable trend. The world of the future will change alongside AI, and we must become the protagonists of that change.
