Chatbots like ChatGPT may not be as intelligent as dogs or cats, according to Yann LeCun, Meta’s chief AI scientist. LeCun received the Turing Award in 2018 for his work in machine learning and made the comments at the Viva Tech conference in Paris. He explained that AI tools trained on large language models are limited because they are only trained on text. While systems like ChatGPT can pass the bar exam for attorneys in the US, they cannot load a dishwasher. LeCun said that AI is missing something big to reach human-level intelligence, insisting that AI cannot even achieve dog-level or cat-level intelligence until something changes.
He also dismissed fears around AI posing a threat to humanity, calling them preposterously ridiculous. LeCun explained that computers will eventually become more intelligent than humans, but that day is still years or decades away. He noted that although AI can learn from massive amounts of text, most of human knowledge has nothing to do with language and that part of the human experience is not captured by AI.
Meta has taken on the challenging task of training AI on video, something that LeCun believes could result in AI systems that are better at understanding the underlying reality of the real world. Until AI can achieve this, the technology will remain limited to text-only applications.
Frequently Asked Questions (FAQs) Related to the Above News
Who made the claim that dogs and cats are smarter than chatbots like ChatGPT?
Yann LeCun, Meta's chief AI scientist, made the claim.
What work did Yann LeCun receive the Turing Award for in 2018?
Yann LeCun received the Turing Award in 2018 for his work in machine learning.
Where did Yann LeCun make the comments about chatbots?
Yann LeCun made the comments about chatbots at the Viva Tech conference in Paris.
What are the limitations of AI tools trained on large language models like ChatGPT?
The limitations of AI tools trained on large language models like ChatGPT are that they are only trained on text.
What tasks can chatbots like ChatGPT do?
Chatbots like ChatGPT can pass the bar exam for attorneys in the US.
Can AI achieve dog-level or cat-level intelligence?
According to Yann LeCun, AI cannot achieve dog-level or cat-level intelligence until something changes.
What did Yann LeCun say about fears around AI posing a threat to humanity?
Yann LeCun dismissed fears around AI posing a threat to humanity, calling them preposterously ridiculous.
Will computers eventually become more intelligent than humans?
According to Yann LeCun, computers will eventually become more intelligent than humans, but that day is still years or decades away.
What is Meta's challenging task in training AI?
Meta's challenging task in training AI is training it on video.
What could result from AI systems being trained on video?
AI systems being trained on video could result in them being better at understanding the underlying reality of the real world.
Please note that the FAQs provided on this page are based on the news article published. While we strive to provide accurate and up-to-date information, it is always recommended to consult relevant authorities or professionals before making any decisions or taking action based on the FAQs or the news article.
