Killing 'delve': We analysed 4,800+ student-authored reports across a decade to uncover how ChatGPT is shaping undergrad writing—style, sentiment, and quality. 📉📈 GenAI isn't just a tool. It's a co-author. 📝 #Preprint: https://t.co/55jb3w8ZqU
Key finding #4 AI policy requiring students to provide a voluntary AI declaration doesn't seem to work, despite clear guidelines and regular communication. We need alternatives!
Key finding #1 ChatGPT-associated stylistic markers (e.g., delve, underscore) surged in prevalence shortly after ChatGPT's launch, but in 2025, we saw a rather consistent decline
"Involvement of episodic memory in language comprehension: Naturalistic comprehension pushes unrelated words closer in semantic space for at least 12h"
📢New paper from: Matthew Mak, Lewis Ball, Alice O'Hagan, Catherine Walsh, & M. Gareth Gaskell
https://t.co/WcpWytCaoC
Am looking to subscribe to some mailing lists for researchers in psycholinguistics, open science, corpus ling, and AI. Used to subscribe to a few but have moved institutes so lost subscriptions... Thanks in advance xx
Research Fellow Opportunity. For informal enquiries, please contact Dr Marina Mendonca (Assistant Professor/Grant Holder) - Research Fellow (109828-1024) at University of Warwick (https://t.co/9AtMbtXRYL)
Have been doing a lot corpus work lately. Struck by the degree of researcher freedom in corpus analysis. Wrote a short article on why I think pre-registration will be beneficial to #corpuslinguistics. Maybe of interest, @CorpusSocialSci? #OpenScience
https://t.co/4jINNNEjyB
1st time lecturing. 200+ students. Man, it's a tough job, standing there talking to mostly blank faces. Felt so rewarding though when there is an occasional nod/smile.
Students who asked questions after lecture and/or left wonderful comments on email/feedback — thank you!
Please share the word about an exciting postdoc opportunity at @warwickpsych
For more information, visit https://t.co/8OkJV5Bwko or email the PI at [email protected]
EPS mid-career prize: An integrated framework for the learning, recognition and interpretation of words.
Gareth Gaskell
https://t.co/ucPI7dAq35
@mggaskell
So apparently this recent paper showing that open science practices can help increase replicability was just retracted by Nature Human Behavior.
Reason appears to be that authors could not find their preregistration?
https://t.co/cOjiK935GR