Over 25 years of serving tech enthusiasts.
Reinforcement learning pioneers criticize the “unsafe state” of AI development. TechSpot is the place to go for tech advice and analysis you can trust.
Who are they? Richard Sutton, Andrew Barto and other pioneers in reinforcement learning are the modern AI models. Sutton, also known as “father of reinforcement learning” is a professor at University of Alberta. Barto is an emeritus professor at the University of Massachusetts. Both scientists aren’t particularly happy with how AI companies apply their life’s research. Richard Sutton
and Andrew Barto won this year’s Turing Award – considered the Nobel Prize of computing – for their significant contributions to machine learning. The two researchers have now spoken out against OpenAI and other AI companies that release potentially dangerous software for end users. They criticized ChatGPT for being a money-making tool that will never create a working artificial intelligence (AGI).
Sutton & Barto developed reinforcement-learning (RL) in the 1980s. They were inspired by behaviorist psychology. Reinforcement learning, along with unsupervised and supervised learning, is one of three basic machine-learning paradigms. Reinforcement learning is a method of teaching AI agents to make the best decisions through trial and errors, similar to what humans do.
OpenAI and Google, among other corporations, build their AI platforms using RL. Financial Times reports that Barto believes that bringing AI software to millions without safeguards is fundamentally wrong. Sutton and Barto used a metaphor to explain that AI companies are building bridges and testing their structural integrity by allowing the public access.
“The idea of having huge data centers and then charging a certain amount to use the software is motivating things, and that is not the motive that I would subscribe to,” .
Companies that are for-profit only look for money-making opportunities. Even if one of them does bring the first AGI to the world, it is merely bragging rights.
AGI proponents believe that superhuman, digital intelligence is close to being a reality and will revolutionize technology. Sutton said that AGI was just a marketing buzzword. Barto said that companies developing AI must gain a better understanding of the human mind before they can responsibly develop systems with human level intelligence.
