The Statistical Society of Australia is having a presentation by Professor Adrian Barnett on the increasing challenges to research quality in the era of large language models.
Date: Wednesday 21 October 2026
Time: 12:00 pm - 1:30 pm AEST
Format: Online via Zoom. Details provided upon registration
From the Abstract:
The pressure to “publish or perish” means that much current research is rushed, with corners cut so that research teams can be first at the cost of being correct. Statistical coding is often rushed and model checking is rarely done. As well as sloppy practice, there is also outright fraud where results are massaged or even entirely faked to get a published paper. These problems are being exacerbated by a research system under strain, as publication numbers are rising and stretching the already limited resource of peer review. AI is adding to the strain, as it allows researchers to quickly write papers and create code in areas that were previously outside their expertise. Large changes are needed to the research system to ensure that published research findings are reliable.