DeepMind Researchers Classify Five Levels of AGI – First Level Achieved

 OpenAI, DeepMind and Meta are all working hard to develop artificial general intelligence - AGI (Artificial General Intelligence) which is a level of artificial intelligence (AI) that is equivalent or better than human intelligence. Earlier this year the CEO of Google DeepMind said that AGI will be achieved in just a few years. This week researchers at DeepMind published five classification levels for AGI with one of them already achieved.

The five proposed AGI level classifications are emerging, competent, expert, virtuoso and superhuman. At this time emerging level AGI has been achieved by ChatGPT and Bard which can perform some tasks as well or better than humans.

To reach the highest level of superhuman, AGI needs to be able to accurately predict the future (such as DeepMind's GraphCast AI weather forecast), talk to animals and read human minds.

The year 2023 is officially the “Year of AI” as dozens of AI models have been announced by various major technology companies. Its use in daily life is also becoming more normal with ChatGPT being used by hundreds of millions of users every week. Will we see AGI in the near future? What will be the impact when an AGI smarter than humans exists?

Previous Post Next Post

Contact Form