Strategic professor of medical education, Australia. Views are my own.
Adelaide, South Australia
Joined February 2012
- Tweets3.4K
- Following506
- Followers886
- Likes891
Lambert Schuwirth retweeted
Future of education 🤔
She got a 1560 on a SAT - with no teacher since the 2nd grade…
She used “a lot of AI apps”
Lambert Schuwirth retweeted
Theories as tools ( ecological)or heuristic( cognitive sci) is the best analogy. Context decides tools, not vice-versa. We shouldn't carry a hammer or we will find everything as a nail.
Lambert Schuwirth retweeted
These #AI detection tools are notoriously inaccurate. You might like to show her teacher this output from GPTZero.
Lambert Schuwirth retweeted
What percentage of students are you willing to falsely accuse of cheating with AI?
There is a trade-off between that and detection rates for AI. At a 10% false positive rate, detectors find 80% or less of AI content. At a 1% rate most find 60% or less.
Don’t trust AI detectors!
Lambert Schuwirth retweeted
"The financial burden of meeting the median APC for publication was 1.94 to 10.26 times higher for authors from lower-income countries than for authors from the US."
This is only one facet of inequalities.
We should stop acting as if each of us has the same opportunity. #MedEd
Lambert Schuwirth retweeted
It is morally wrong to use AI detectors when they produce false positives that smear students in ways that hurt them and where they can never prove their innocence.
Do not use them. bloomberg.com/news/features/…
Lambert Schuwirth retweeted
Am featured alongside @shaananc in this piece on AI detectors. Nothing new from me, but it gets the message out
Victorian teachers are being warned not to rely on artificial technology detection programs, after a number of uni students were wrongly accused of cheating on assignments > bit.ly/46jUrDj
Lambert Schuwirth retweeted
What's wrong with student evaluations of teaching (SETs)? Well, honestly, a sh@%-ton. But this little figure tells a large part of the story of why they're so problematic and the consequences of their (over)use in #HigherEd. BTW, the figure comes from this AWESOME #OpenAccess #OA article... #MedEd
tandfonline.com/doi/full/10.…
Lambert Schuwirth retweeted
TLDR - no
It’s time to re-examine single episode high stakes exams
Change in assessment *is* coming.
You can either engage, listen and control the future narrative or not and…see where that gets you.
#MedEd #medtwitter
Lambert Schuwirth retweeted
TEQSA has issued an information request to all registered #HigherEd providers, asking for an action plan addressing the risk #GenAI poses to award integrity.
teqsa.gov.au/about-us/news-a…
ALT Image has a blue background and white writing which reads "Information request: AI risk mitigation". The TEQSA logo is in the top-left. In the bottom-left is an orange circle with an illustration of paper and, underneath the illustration, an arrow pointing down. The bottom-right corner of the image is orange with TEQSA in blue writing.
Lambert Schuwirth retweeted
Was your abstract rejected for #AMEE2024? Ours too.
We have great news! Now you can share your work with a global audience.
We (me & @DrPriyaKhanna) created a YouTube channel: Rejected but Respected
Submit your rejected works!
The only criterion: Being rejected by AMEE 2024
Lambert Schuwirth retweeted
A thing that I keep seeing in organizations I talk to: a few senior people are experimenting with AI in their work and realizing that this is going to be a huge deal in their industry.... yet they have huge trouble getting their colleagues to even try the systems out seriously.
Lambert Schuwirth retweeted
Selamat Hari Raya Aidilfitri and Eid Mubarak to colleagues and friends celebrating near and far! May peace, good health and happiness surround you and your loved ones always.#Eid Mubarak #NCL @NUMedMalaysia
Lambert Schuwirth retweeted
The problem isn't that we've created assessments that ChatGPT can pass, but that our assessments and interpretive paradigms are still not rich enough to reflect the complexity of performance or the process of learning.
Lambert Schuwirth retweeted
Goodhart's Law: “When a measure becomes a target, it ceases to be a good measure.” A famous historical example is the "cobra effect": in India during British rule, the government, concerned about the number of venomous cobras, offered a bounty for every dead cobra. Initially, this was successful. Then people started breeding cobras for the income. The reward programme was scrapped so the breeders set their snakes free & increased the cobra numbers.
Two strategies to "cobra-proof" our goals & metrics:
1) Define metrics clearly in ways that reduce the potential for them to be misinterpreted or gamed
2) Use a diverse set of metrics collectively to measure the “true goal”
medium.com/illumination/the-…. By @gauravjainio
Lambert Schuwirth retweeted
You know what the biggest problem with pushing all-things-AI is? Wrong direction.
I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes.
Lambert Schuwirth retweeted
I said in the newspaper that AI detection research assumes users are idiots who don't try to not get caught. Well here's a *preprint* (ie not yet peer reviewed) paper where they find AI detection much less accurate when attempts are made to foil detection arxiv.org/abs/2403.19148
