We emphasize the need for diverse samples, method validation, transparency, and ethical considerations to ensure robust, generalizable findings free from demographic biases. We advocate for rigor and replicability over convenience in integrating AI into psych research.
15.11.2023 23:50 — 👍 1 🔁 0 💬 0 📌 0
We further demonstrate that smaller, simpler, more transparent, and importantly reproducible methods can often match or surpass the performance of these large complex models.
15.11.2023 23:50 — 👍 1 🔁 0 💬 2 📌 0
We introduce the concept of "GPTology" – the rapid and uncritical use of Large Language Models (LLMs) in psychology, which risks producing misleading research due to biases and opaque methodologies.
15.11.2023 23:50 — 👍 0 🔁 0 💬 1 📌 0
We investigate the application of LLMs in psychological research, addressing both their applications in text analysis and as stand-ins for human participants.
15.11.2023 23:49 — 👍 0 🔁 0 💬 1 📌 0
New Preprint: "Perils and Opportunities in Using Large Language Models in Psychological Research." osf.io/tg79n
@mdehghani.bsky.social @aliomrani.bsky.social @mohammadatari.bsky.social
15.11.2023 23:48 — 👍 6 🔁 0 💬 1 📌 1
