Nvidia's CEO foresees Artificial Intelligence passing human-level tests within five years

Nvidia's CEO, Jensen Huang, projected on Friday that artificial general intelligence (AGI) might reach the capability to pass human-level tests within the next five years, under certain definitions of AGI. Speaking at an economic forum at Stanford University, Huang addressed the timeline for achieving AGI, a long-standing ambition in Silicon Valley to develop computers with human-like cognitive abilities.

Huang emphasized that the timeframe for AGI's arrival heavily depends on its definition. "The question of when we will achieve artificial general intelligence hinges on how we define its goals. If we measure it by the ability to pass a variety of human tests, we're looking at a near future where AGI becomes a reality," Huang remarked.

He illustrated his point by stating:

If I gave an AI every single test you can imagine, and those tests were presented to the computer science industry, I believe that in five years, we'll excel in each one.

This statement came as Nvidia, the leading manufacturer of AI chips for systems such as OpenAI's ChatGPT, celebrated reaching a $2 trillion market value.

Currently, AI systems have succeeded in passing complex exams, including legal bar exams, yet they still face challenges with more specialized fields, like gastroenterology. Huang is optimistic that within five years, these hurdles will be overcome, allowing AI to pass any test presented to it.

However, Huang also noted that reaching AGI might take much longer by other standards, given the ongoing debate among scientists over the precise workings of the human mind.

Achieving AGI is challenging from an engineering perspective without a clear understanding of how to replicate human cognition.

In response to inquiries about the need for more chip factories to support AI's growth, with OpenAI's CEO Sam Altman highlighting the requirement for additional facilities, Huang acknowledged the necessity but pointed out that advancements in chip technology and AI algorithms would also play a crucial role in meeting demand. "While the need for more fabs is undeniable, it's essential to remember that the efficiency of AI processing and algorithms is improving exponentially, potentially offsetting the sheer volume of chips required," Huang concluded.