LLMs still fragile in complex reasoning tasks: Cognizant AI chief

Cognizant's Chief AI Officer Babak Hodjat highlighted the challenge of trusting large language models (LLMs) for complex business operations. He explained that LLMs often fail in long reasoning chains, necessitating human oversight and robust evaluation frameworks. Cognizant is developing methods like semantic density and real-time governance to enhance AI reliability and mitigate risks.

LLMs still fragile in complex reasoning tasks: Cognizant AI chief
Cognizant's Chief AI Officer Babak Hodjat highlighted the challenge of trusting large language models (LLMs) for complex business operations. He explained that LLMs often fail in long reasoning chains, necessitating human oversight and robust evaluation frameworks. Cognizant is developing methods like semantic density and real-time governance to enhance AI reliability and mitigate risks.