Content
summary Summary

DeepMind co-founder Demis Hassabis believes that current AI capabilities are overrated. However, he believes the technology's long-term potential, especially when it comes to artificial general intelligence (AGI), is still underestimated.

Ad

In a recent podcast, Hassabis described the public perception of AI as a double-edged sword. The widespread interest shows that AI is poised to have a positive impact on everyday life. But this attention also brings noise to the field. For AI veterans like Hassabis, it's surreal to watch the public slowly grasp the future importance of AI.

Although he believes current AI models are "hyped too much," Hassabis is impressed by their ability to infer real-world information despite learning only from language. "They're learning in an abstract domain, so it's pretty surprising they can infer some things about the real world from that," Hassabis said.

Short-term hype vs. long-term potential

Hassabis believes AI's capabilities are overestimated in the short term. Many people and startups claim the technology can do things it can't do yet, he says. Hassabis calls them "crazy ideas" that venture capitalists are chasing even though the underlying technology doesn't exist.

Ad
Ad

However, Hassabis believes AI's long-term potential is still underestimated, especially regarding AGI and beyond. He thinks AGI could be achieved within the next decade. DeepMind was conceived as a 20-year project, with that timeframe ending around 2030.

"I wouldn't be surprised if it [AGI] comes in the next decade," says Hassabis, reiterating that he sees enormous potential for humanity through AGI, including curing diseases, developing clean energy, and solving climate change.

Hassabis also stressed the need to address the risks of human misuse and uncontrollable AI systems as these systems become more capable. He calls for international cooperation in regulating AI, predicting that the technology will be difficult to contain within national borders. In the final stages of AGI development, Hassabis envisions a scientific "Avengers" group of the world's top AI researchers working together.

Ad
Ad
Join our community
Join the DECODER community on Discord, Reddit or Twitter - we can't wait to meet you.
Support our independent, free-access reporting. Any contribution helps and secures our future. Support now:
Bank transfer
Summary
  • Demis Hassabis, co-founder and CEO of Deepmind, believes that the capabilities of current AI models are overstated. Many claim that the technology can do things that it cannot yet do, and that VCs are chasing "crazy ideas" that are not yet technologically feasible.
  • In the long term, Hassabis says that the potential of AI is still underestimated, especially when it comes to Artificial General Intelligence (AGI). Hassabis believes it is possible to achieve AGI in the next decade and sees enormous potential for humanity, for example in curing diseases and solving the climate problem.
  • At the same time, however, the risks of human abuse and uncontrollable AI systems must be contained. The Deepmind boss advocates international cooperation in regulating AI and would like to see a scientific "Avengers" team of the world's best AI researchers in the final stages of AGI development.
Sources
Online journalist Matthias is the co-founder and publisher of THE DECODER. He believes that artificial intelligence will fundamentally change the relationship between humans and computers.
Join our community
Join the DECODER community on Discord, Reddit or Twitter - we can't wait to meet you.