'Godfather Of AI' Explains Why ChatGPT, Copilot, And Other AI Chatbots Hallucinate
Portfolio Pulse from Rounak Jain
'Godfather of AI' Yann LeCun, Meta Platforms Inc.'s chief AI scientist, explained on the Lex Fridman podcast why AI chatbots like OpenAI's ChatGPT and Microsoft Corp.'s Copilot sometimes 'hallucinate' or generate nonsensical responses. He attributed this to the autoregressive prediction model used by these chatbots, where the probability of generating an irrelevant answer increases with each word produced. This phenomenon occurs especially when the chatbot encounters questions outside its training scope.

March 08, 2024 | 2:24 pm
News sentiment analysis
Sort by:
Ascending
NEUTRAL IMPACT
Yann LeCun, Meta Platforms Inc.'s chief AI scientist, discussed AI chatbots' tendency to 'hallucinate', shedding light on the challenges faced by AI technologies.
While the article focuses on the explanation of AI chatbots' behavior, it indirectly showcases Meta's involvement and expertise in AI research, which could reassure investors about Meta's capabilities in AI but doesn't directly impact its financial performance in the short term.
CONFIDENCE 90
IMPORTANCE 60
RELEVANCE 70
NEUTRAL IMPACT
Microsoft Corp.'s AI chatbot Copilot was mentioned as an example of AI 'hallucination', a challenge in the development of AI technologies.
The mention of Microsoft's Copilot as an example of AI 'hallucination' highlights the challenges in AI development but doesn't directly affect Microsoft's stock in the short term. It shows Microsoft is at the forefront of AI innovation, which could be seen positively by investors.
CONFIDENCE 90
IMPORTANCE 60
RELEVANCE 70