AI chatbots may not match human intelligence levels, says Meta’s AI chief

Photo of author


Yann LeCun, Meta’s chief AI scientist, expressed his views that AI chatbots like ChatGPT and Gemini, which are powered by large language models (LLMs), will not be able to equal humans in the areas of reasoning and planning.

Lok Sabha Election Results 2024

During an interview with the Financial Times, LeCun highlighted that large language models currently lack a strong grasp of logic and have limited understanding. He stated that these models do not possess essential capabilities like understanding the physical world, maintaining memory, reasoning in a logical manner, or hierarchical planning.

According to LeCun, large language models such as ChatGPT pose inherent risks because their accuracy in responses is heavily reliant on the quality of the training data. He explained that these models primarily leverage accumulated knowledge from extensive training data rather than true reasoning. Despite these limitations, LeCun acknowledged the practical utility of LLMs like ChatGPT and Gemini.

When discussing the path to achieving human-level intelligence in AI, LeCun revealed that Meta’s Fundamental AI Research (Fair) lab, comprising approximately 500 experts, is actively exploring a new AI system capable of acquiring common sense and understanding the world. This strategy, termed ‘world modeling,’ carries potential risks for Meta, as investors often seek rapid returns on AI investments.

Festive offer
Recently, Mark Zuckerberg announced plans to increase AI investments at Meta in a bid to establish the company as a global AI leader. Following this announcement, the social media giant experienced a decline of around $200 billion in valuation.


Source link

Leave a Comment

For security, use of Google's reCAPTCHA service is required which is subject to the Google Privacy Policy and Terms of Use.

Share to...