ChatGPT doesn't think like a human, so why does it speak like one?
Gaskell, G.
(19 September 2025)
ChatGPT doesn't think like a human, so why does it speak like one?
Impact of Social Sciences Blog.
Large language models like ChatGPT adopt aspects of human speech to give the impression of human intelligence. George Gaskell highlights the limits of ChatGPT’s conversational ethics, common sense and cultural sensitivity and suggests this kind of ‘human talk’ leaves much to be desired for users.
| Item Type | Blog post |
|---|---|
| Copyright holders | © 2025 The Author(s) |
| Departments | LSE > Academic Departments > Methodology |
| Date Deposited | 06 Nov 2025 |
| URI | https://researchonline.lse.ac.uk/id/eprint/129745 |
Explore Further
ORCID: https://orcid.org/0000-0001-6135-9496