New York: In response to recent concerns raised by Elon Musk regarding AI’s impact on employment, Yann LeCun, Meta’s AI chief, weighed in on the discussion, asserting that artificial intelligence cannot attain human-like intelligence. LeCun’s statements shed light on the limitations of AI systems, particularly in understanding and reasoning complex human interactions.
LeCun emphasized that large language models (LLMs) like ChatGPT are inherently incapable of achieving human-level intelligence due to their inability to reason effectively. These remarks from Meta’s AI chief come at a time when major tech players such as Google and OpenAI are heavily investing in AI research and development.
While the aspiration of achieving Artificial General Intelligence (AGI) remains a goal for many in the industry, LeCun expressed skepticism about the current trajectory. He highlighted the vast disparity between AI systems and biological intelligence, noting that animals and humans exhibit remarkable intelligence with significantly less training data compared to LLMs.
“Animals and humans get very smart very quickly with vastly smaller amounts of training data than current AI systems. Current LLMs are trained on text that would take 20,000 years for a human to read,” LeCun remarked, underscoring the inefficiency of existing AI models in replicating human-like intelligence.
According to LeCun, reliance on current LLMs for achieving human-level intelligence is misguided, as these models require extensive and precise training data to produce accurate responses to human prompts. He cautioned against overstating the capabilities of AI systems, highlighting their intrinsic limitations and potential safety concerns.
Moreover, LeCun challenged the notion that AI models could rival the intelligence of animals, citing the remarkable cognitive abilities demonstrated by corvids, parrots, dogs, and octopuses with far fewer neurons and parameters than current AI systems.
Contrary to Elon Musk’s grim prognosis of AI-induced job displacement, LeCun asserted that today’s AI models are fundamentally different from human cognition and lack the nuanced understanding and adaptability inherent in biological intelligence. He cautioned against equating AI’s capabilities with those of humans or animals, emphasizing the need for a more nuanced understanding of AI’s limitations and potential applications.
