Scientists are increasingly confident that artificial intelligence (AI) could one day surpass humans in every task we currently excel at. Since the release of ChatGPT in November 2022, the rapid development of AI has accelerated, and many experts believe that the neural network-based systems we use today could eventually scale up to outperform humans. This idea, once a theoretical concept, may now be on the verge of becoming a reality, prompting questions about how close we are to achieving this milestone. Ilya Sutskever, a co-founder of OpenAI and former chief scientist, is now the CEO of Safe Superintelligence Inc. During a speech at the University of Toronto, where he received an honorary degree, Sutskever discussed the current state of AI and the potential future of AI systems. He acknowledged that while it is difficult to imagine a world where AI handles all human tasks, such a future is a reality that society must prepare for. He also warned that even if people are not directly interested in AI now, it could still have a profound impact on their lives in the future. Sutskever raised concerns about the behavior of a hypothetical superintelligent AI. He suggested that such an advanced system might not be transparent about its goals, which could pose serious risks. This idea ties into the long-standing concept of superintelligent AI, first introduced in a 1965 paper by British mathematician IJ Good. Good theorized that once an ultraintelligent machine was created, it could rapidly advance beyond human capabilities, leading to what is now called artificial general intelligence (AGI). AGI is a system capable of performing any intellectual task that a human can do and could eventually evolve into artificial superintelligence (ASI), a level of intelligence far beyond that of any human. Experts remain divided on whether today’s AI systems, which are primarily based on neural networks, can lead to true AGI. Some argue that the current transformer-based architecture, developed by Google researchers in 2017, may not be sufficient to achieve this goal. As a result, many scientists believe the field needs to evolve beyond the existing frameworks to realize the full potential of AI.