Sign In  |  Register  |  About Mill Valley  |  Contact Us

Mill Valley, CA
September 01, 2020 1:29pm
7-Day Forecast | Traffic
  • Search Hotels in Mill Valley

  • ROOMS:

Tech CEO's ambitious 5-year timeline for 'competitive' AI comes with caveats, experts argue

Experts argue that the development of AI to achieve AGI relies on datasets available and how broadly researchers want to define the achievements of AGI models.

Nvidia CEO Jensen Huang predicted that artificial intelligence (AI) will achieve thinking and learning at a "competitive" level to human beings within five years, but experts told FOX Business that such an ambitious timeline comes with many caveats. 

"Broadly, I think it’s really hard to gauge the pace of development, but just over the past few years the gains have been really impressive with respect to generating more and more realistic outputs," Valerie Wirtschafter, a Brookings fellow in Foreign Policy and the Artificial Intelligence and Emerging Technology Initiative, explained. 

Huang, speaking Wednesday at the New York Times DealBook Summit, argued that achieving artificial general intelligence (AGI) came down partially to how researchers and developers define the term, but that in some cases it remains a realistic goal within the next decade – or sooner. 

"If we define AGI as a piece of software – a computer that can take a whole bunch of tests and these tests reflect basic intelligence — and by completing those tests deliver results fairly competitive to a normal human, I would say that within the next five years, you can see AIs that can achieve those tests."


Nvidia has assumed a key role in the AI development rush due to the high demand for chips and Graphics Processing Units (GPUs), both of which are essential components for advanced AI platforms and models to achieve their high computing processing potential. 

The company now dominates the market for chips used in AI applications, according to Barrons: Nvidia’s revenue rose 206% over the prior year in its latest quarter thanks to the surge in AI interest and demands. 

The United Kingdom pledged to spend hundreds of millions of pounds on chips to help boost its own AI development ambitions, with Nvidia the primary beneficiary of that expenditure. The U.S. currently retains a heavy reliance on Asia for its chip supply – primarily Taiwan, which dominates the chip production market. 


Chips alone will not determine how quickly countries and companies alike will be able to achieve major AI breakthroughs, such as AGI, which has proven the chief goal of many AI researchers as it would prove the point at which AI has achieved seemingly human intelligence.

A 2020 report from consulting giant McKinsey said a true AGI would need to master skills like sensory perception, fine motor skills and natural language understanding – an AI that can "do anything and maybe as well or better than a human," according to one expert. 

OpenAI allegedly had developed an algorithm, title Q*, that researchers at the company believed could lead to true AGI, which the company defines as systems that "surpass humans in most economically valuable tasks," Reuters reported. The model could solve "certain mathematical problems," which made researchers "optimistic" about the program's future success. 

Wirtschafter argued that some AI systems have already proven "competitive" in certain respects, pointing to Meta’s Cicero, which "excelled in the game of Diplomacy well beyond the average human." 


"Some of these things are niche, and are fine-tuned for specific tasks, like a game of Diplomacy. I suspect with more fine-tuning on specific tasks, we’ll see other AI gains too, but that’s not the same thing as artificial general intelligence, which requires the ability to reason and problem solve in a way people do," Wirtschafter noted. 

"The biggest needs for rapid improvements in output quality are always the availability of bigger datasets, more computing power, and increased investments, and of course advances in deep learning and frontier models," Writschafter added. 

Emily de La Bruyere, a senior fellow at the Foundation for Defense of Democracies specializing on China and technology, focused on how different actors – such as China and the U.S., the leading AI nations – place their "bets" and initiatives as significant factors in the timeline for AGI development. 

"China's bets are largely… in industrial AI and autonomous driving, as opposed to an AI that’s supposed to write the Great American Novel or a more humanoid role, and I think there's a difference in those and the feasibility of them panning out," de La Bruyere explained, arguing that China’s goals are "a little more realistic and also a little bit more impactful."


Experts have started to worry, however, that the flood of AI-generated material could negatively impact the timeline of development as the technology faces the problem of potentially self-cannibalizing thanks to an arguably corrupted source data set. 

"I think being able to pass on the volume of content is a challenge, and I think, from the beginning of even earlier waves of hype, I think learning and data sets have always been a limiting factor," Nathan Picarsic, also a senior fellow at FDD focusing on China and technology, said, stressing that the lack of quality control on the datasets fed into AI for training can lead to "recurring learning" through the systems. 


"I think it’s going to be hard for applications to have higher quality learning and training data," he added. "Obviously, having a ton of content flooding channels that we might use for other purposes is going to be a challenge for society and for policy makers alike." 

Fox News Digital’s Kelsey Koberg contributed to this report. 

Data & News supplied by
Stock quotes supplied by Barchart
Quotes delayed at least 20 minutes.
By accessing this page, you agree to the following
Privacy Policy and Terms and Conditions.
Copyright © 2010-2020 & California Media Partners, LLC. All rights reserved.