Students at Yale and other universities are now using AI chatbots in real-time during class discussions, typing professors' questions directly into ChatGPT and reading the responses aloud as their own contributions. Amanda, a Yale student, watched a classmate "typing ferociously" during an awkward silence, feeding the professor's question about assigned reading into an AI tool. Another student, Jessica, reported seeing "every single person putting every single PDF into AI" at the start of each class.
This builds on research I covered showing most AI users abandon critical thinking when using these tools. What's striking here is the speed of adoption—students aren't just using AI for homework anymore, they're using it as a real-time crutch during live academic discussions. The result is predictably flat: "Everyone now kind of sounds the same," Amanda noted, contrasting sharply with her freshman year when classmates "approached from different angles and offered different commentary."
A broader survey by the National Education Union in England found 66% of secondary school teachers reporting declining critical thinking skills among students using AI, more than double the rate among primary teachers. USC psychology professor Morteza Dehghani, who co-authored recent research on AI's cognitive effects, called the implications "quite scary," warning that intellectual homogenization could "affect our society greatly." The survey also revealed that students no longer feel the need to learn spelling due to voice-to-text technology.
For developers building AI tools, this raises serious questions about real-time usage patterns we might not be tracking. If students are using our APIs during live conversations, we're essentially automating human discourse without realizing it. The classroom is becoming an unintentional testing ground for AI-human collaboration—and early results suggest we're not designing these interactions thoughtfully enough.