Abstract:
"At the time of writing, GPT-4 has been available for
less than 48 hours and already we know that it can
pass the majority of standardized tests as well as
the exams of several professional orders (OpenAI,
2023). Compared to its predecessor GPT-3.5, released four months earlier in November 2022, the
new version does better on the vast majority of
tests and scores up to two to three times higher. We can only speculate on what capabilities generative
artificial intelligence tools (GAIs) will have come
fall 2023. Do we really want to embark on a technological arms race against Silicon Valley? What if
our safe-conduct were pedagogical? In this article,
I want to approach the phenomenon of GAIs from
a pedagogical perspective, particularly with regard
to the thorny issue of assessment validity." -- AQPC
