Neural network study sheds light on how AI chatbots 'reason' like humans

A Tech Xplore study found that AI chatbots do not think symbolically like humans, yet can offer cogent writing and math advice when prompted. The work sheds light on how large language models 'reason'.

• Humans think symbolically using numbers, variables and words
• AI chatbots do not use symbolic thinking
• When prompted, models give coherent math and writing advice