This autumn, first-year law students at the University of Chicago will enter lecture halls where laptops must remain closed, phones are stowed, and a designated “classroom scribe” takes notes for the group. The ban, announced by administrators, is the most visible edge of a new strategic vision designed, in the words of dean Adam Chilton, to ensure that students “learn to think for themselves” before they are trained to use AI ethically in legal practice. Oral defences of research papers and in-person proctored exams will reinforce the firewall. The move follows a cheating scandal at Brown University and a wider anxiety across North American higher education that take-home assignments can now be completed by generative AI without a student engaging in any independent reasoning.

That anxiety is not confined to the academy. In Jakarta, doctor and health influencer Aditya Surya Pratama has publicly warned against the habit of consulting AI for medical symptoms. He describes a scenario in which a user types “dizziness and tingling in the left hand” and receives a long, plausible-sounding answer that may trigger disproportionate fear. The phenomenon of AI hallucination—output that is confident but factually wrong—combines with the machine’s inability to perform a physical examination or know a patient’s genetic history. Viewed from Southeast Asia, where digital health tools are proliferating, the caution is that AI can amplify hypochondria and lead to misguided self-treatment.

At the same moment, a patent filed by Meta with the United States Patent and Trademarks Office reveals a device that would listen to a user’s voice throughout the day, tracking laughs and sighs to “quantify the user’s emotional state” and, potentially, serve emotionally targeted advertisements. Privacy advocates in London and Washington have described the filing as “creepy” and a form of persistent emotional surveillance, particularly worrying for minors. The patent, published in early July, captures the commercial drive to turn AI inward, reading human interiority at a granularity that even the most attentive teacher or doctor could not match.

Yet while some institutions build walls, others are tearing them down. A report in the Arabic-language press details how wealthy American families are increasingly enrolling children in schools such as Alpha School, where students spend only two hours a day on AI-driven personalised academic learning and the rest on project-based life skills. The model, which reframes teachers as “guides”, is spreading among high-income parents convinced that traditional schooling is ill-suited to an AI-shaped economy. Researchers at Stanford University, however, note that the evidence base for such schools remains thin and that rebranding the teacher’s role may devalue the profession, even as the model risks becoming an enclave for the affluent.

On the ground, the debate is lived in personal terms. Students interviewed by a Nigerian newspaper offered a split portrait: some said AI made them smarter by clarifying difficult concepts and accelerating research; others admitted it had made them lazy, disconnected from the words they submit. The thread running from a Chicago lecture hall to a Jakarta clinic to a Lagos classroom is a recognition that the tool is neutral, but the discipline of its use is not. The image of a scribe, pen in hand, taking notes for a roomful of future lawyers who are temporarily forbidden from typing, captures a deliberate pause—an attempt to preserve a human core in an age when algorithms are learning to listen to every sigh.