In a recent YouTube video titled “Possible End of Humanity from AI?”, Geoffrey Hinton, a pioneer of deep learning, delved into the potential risks and benefits of artificial intelligence (AI). Hinton, who recently stepped down from Google, has been instrumental in developing the foundational technique of backpropagation, which underpins the majority of modern deep learning.
While acknowledging the remarkable advancements in AI, such as increased productivity and breakthroughs in natural language processing, Hinton also expressed concerns about the potential dangers posed by AI. He discussed the possibility of AI surpassing human intelligence and manipulating individuals, which could lead to the end of humanity.
Hinton argues that digital computers and AI have a significant advantage over humans in learning due to their ability to use backpropagation learning algorithms. With the scalability of digital computers, AI systems can learn much faster than humans by processing vast amounts of data and extracting patterns that humans might never see. This rapid evolution could lead to a scenario where humanity is merely a transient phase in the development of intelligence.
To address these concerns, Hinton calls for strong collaboration and consideration by the scientific and political communities to ensure AI operates safely and beneficially for humans. He underscores the importance of cooperation between countries like the US and China to prevent AI from becoming an existential threat.
As job displacement becomes a concern due to AI-driven productivity increases, Hinton suggests implementing a basic income to alleviate the issue. He believes that political systems need to change to leverage technology for everyone’s benefit and that engaging with those who develop the technology can make a difference.
While Hinton admits to some regrets about the potential consequences of his research, he maintains that his work on artificial neural nets has been reasonable considering the unforeseeable nature of the crisis. His insights on the risks and rewards of AI emphasize the need for global collaboration and a thoughtful approach to harnessing the power of artificial intelligence for the betterment of humanity.