Three years ago, researchers tested AI on real Australian law exams and it struggled. They wanted to know if this is still the case.
- Senior Lecturer in Law and Emerging Technologies, Western Sydney University
In stress-testing AI models, it’s not hard to push them to the brink and make them threaten to harm humans.
Researchers developed an educational chatbot to help law students and save time. But the AI kept giving inaccurate, misleading and even incorrect feedback.
A recent study shows generative AI is not the academic powerhouse it is often made out to be.
There won’t be an easy tech fix for the questions about authorship raised by ChatGPT and other text generators.
Contact Armin for
- General
- Media request
- Speaking request
- Consulting / Advising
- Research collaboration
- Research supervision
- Location: Sydney, Australia
- Website
- LinkedIn: https://www.linkedin.com/in/armin-alimardani/
- X (formerly Twitter): @RminAlimardani
- Article Feed
- ORCID
- Joined
