[News Terms] Musk: 'AGI' Could Emerge Next Year

Movie 'Iron Man' 'J.A.R.V.I.S.' Becoming Reality?
Jensen Huang & Sam Altman Say "Within 5 Years," Faster Than Expected

'AGI (Artificial General Intelligence)' refers to an artificial intelligence system capable of performing a wide range of intellectual tasks at a level similar to that of humans.

Currently, artificial intelligence (AI) is specialized and optimized for specific tasks and cannot adapt to other tasks. However, AGI possesses the ability to flexibly adapt to various tasks, enabling it to respond not only to specific problems but to all given situations. Therefore, it is also called 'strong AI,' 'artificial general intelligence,' or 'general-purpose AI.'

AGI is one of the ultimate goals of AI research. Since it aims to develop machines with intelligence surpassing human levels, it is expected to have learning abilities, reasoning skills, problem-solving capabilities, and creativity similar to those of humans.

It is a concept similar to the AI system 'J.A.R.V.I.S.' featured in the movie 'Iron Man.' It can autonomously manage schedules and provide advice like a human, representing the next generation of AI. In other words, if AGI is developed and becomes widespread, it would be like having a personal assistant for each individual, which is expected to bring significant changes to daily life.

Because of this, there is ongoing controversy about the actual development of AGI. Professor Geoffrey Hinton of the University of Toronto, Canada, warned that "the commercialization of AGI will happen faster than expected, and the associated risks are significant." He argued that it could materialize in the form of killer robots like those in the movie 'Terminator,' posing a real threat to human survival.

Dr. Stephen Hawking and Elon Musk have also expressed concerns that the emergence of AGI could replace humanity due to biological limitations.

On the other hand, some experts argue that the development of AGI is impossible. Professor Andrew Ng of Stanford University and Professor Choi Yejin of the University of Washington claim that AI finds it difficult to understand human common sense and that practical constraints such as data scarcity make AGI development challenging.

However, everyone agrees on the risks of AGI technology and the need for control measures. There have been calls for 'alignment' between humans and AGI to prevent and control risks. Consequently, discussions on AI regulation, the enactment of laws, and the establishment of independent supervisory bodies are also underway.

Elon Musk, CEO of Tesla, has attracted attention by stating that AGI could emerge as early as next year. In an interview on the 8th (local time) with Nikolai Tangen, CEO of the Norwegian sovereign wealth fund, conducted on X (formerly Twitter), Musk said, "If AGI is defined as AI smarter than the smartest human, it could probably happen next year, or within two years, for example."

Musk used expressions such as 'probably' and 'think' and did not provide concrete evidence, but his remarks are noteworthy given that he is also developing AI models through his AI startup xAI.

His forecast significantly precedes that of Jensen Huang, CEO of Nvidia, a leading AI chip company, who predicted that AGI at a human level would appear within five years. Sam Altman, CEO of OpenAI, the developer of ChatGPT, also expects AGI to be built within the next four to five years.

Shane Legg, co-founder of Google DeepMind and chief AGI scientist, estimated a 50% chance of AGI development by 2028. Dario Amodei, CEO of AI startup Anthropic, predicted an even faster timeline, expecting AI capable of human-level performance within two to three years.

On the other hand, Yann LeCun, Meta’s Vice President and Chief AI Scientist, considered one of the 'Big Four' in AI, said, "Human-level AI will come someday, but not within three or five years," adding, "The timing is hard to predict, and it may take longer than we expect."