A leading scientist in the world has stated that it is possible to build an AI (artificial intelligence) that exhibits general intelligence similar to humans by 2027. This could open the door to “superintelligence” in a very short time.
Ben Goertzel, a computer scientist and CEO of SingularityNET, made this statement during his closing speech at the AGI 2024 Summit in Panama City, Panama. He is dubbed the “father of AGI” after helping popularize the term Artificial General Intelligence (AGI) in the early 2000s.
Ben Goertzel believes we are on track to create human-level intelligent AI by 2027. (Photo: SingularityNET)
The best AI systems currently deployed are considered “narrow AI” as they may outperform humans in specific fields based on training data, but cannot surpass human capabilities overall. These narrow AI systems, ranging from machine learning algorithms to large language models (LLMs) like ChatGPT, struggle with reasoning and understanding context like humans.
However, Goertzel notes that AI research is entering a phase of exponential growth, and there is evidence that Artificial General Intelligence (AGI) – where AI can match human capabilities in several areas independent of its original training data – is within reach.
Goertzel suggests that 2029 or 2030 may be the most appropriate years for humanity to create the first AGI agent, but it could happen as early as 2027.
If such an agent is designed to have access to and rewrite its own code, it could quickly evolve into Artificial Superintelligence (ASI) – which Goertzel defines as an AI with cognitive and computational capacities surpassing that of all human civilization.
Goertzel stated: “No one has yet created artificial intelligence at the human level. It is reasonable to expect that we could achieve AGI at the human level within the next 3 to 8 years.”
He pointed to three pieces of evidence to support his argument, the first being the modeling by computer scientist Ray Kurzweil in his book The Singularity is Near (Viking USA, 2005), which has been refined in his forthcoming book The Singularity is Nearer (Bodley Head, June 2024).
In his book, Kurzweil developed predictive models showing that AGI could be achieved by 2029, primarily focusing on the exponential nature of technological growth in various fields.
Goertzel also noted the improvements made to LLMs in recent years, which have awakened many people worldwide to the potential of AI.
According to Goertzel, the third piece of evidence lies in the work on building the “OpenCog Hyperon” infrastructure, along with related software systems and the upcoming AGI programming language named “MeTTa”, to support it.
OpenCog Hyperon is a form of AI infrastructure that involves the integration of existing and new AI models, including LLMs as a component.
However, Goertzel admits that he could be wrong and that a quantum computer with a million qubits or something similar may be necessary.
Goertzel added: “Once we reach human-level AGI, it could lead to fully transcendent AGI within a few years – unless AGI threatens to stifle its own development due to conservatism.”
