If there is a chance that AI will touch their references, then they need an AI-free automated reference checker. Here's one: https://www.refrunner.com/
Was thinking about this myself. As a libertarian I don't want regulation exactly but I do think we need a legal framework for handling how agents can act in the market. Like can an AI agent have "personhood" ? Or are the people they are agents for the responsible party? Or are the companies that developed the AI models? Who owns liability if an agent does something harmful?
Is that the sound of the French government fleeing south and the thud of Wehrmacht boots tramping through the defenseless streets of Paris?
Like, c'mon! How hard is it to say, "Any paper that scores 50% AI or more on Pangram is an automatic fail and any paper in the 25-49% AI range will be subject to additional critique?" Pangram may not be accurate at a sentence by sentence level, but it is extremely accurate at judging large blocks of text (700 words or more). Correct me if I'm wrong, but I thought it was the business of educators to teach students how to think for themselves. That includes writing and the excruciating articulation of ideas which is really a student's first test of them, manifesting them into sentences and connecting concepts with evidence in a step-by-step process. But this "AI encouragement" strikes me more as an abandonment of that process, much as the French government abandoned the people of Paris instead of fighting block by block. A student's capacity for critical thinking deserves our attention and our defense! But now it seems we are granting permission to students to outsource their primary job to an unintelligent, probability calculator... unless, of course, we have only every cynically defined that job as submitting a certain quantity of assignments and checking a certain order of boxes on the way to receiving a perfunctory participation grade.
It used to be that plagiarism was the thing that plagued professors. Now it's AI. Different foe, but same principle. And much like plagiarism, after a while, you don't need even need Pangram to discern a lot of AI writing. I'm glad my professors fought for academic standards that help students reach and grow. It would be a shame if the evolution of technology causes the next generation of profs to throw up their hands in futility just because they don't want to have a couple of difficult conversations.
Pragmatism is a virtue, up to a point. But let's not delude ourselves for what this is, wholesale surrender thinly veiled as fairness.
The requirement of at least one responsible human in the loop should be part of every automated system. The old "computer error" excuse should never have been tolerated.
If there is a chance that AI will touch their references, then they need an AI-free automated reference checker. Here's one: https://www.refrunner.com/
hadn't run into that one, thanks!
This is very good, and similar to what I am doing in my two classes (mine is a bit simpler). I might do a post on this later, linking to yours.
honestly, I think we really need a lot of these kinds of discussions out there, if just to help our junior colleague navigate this shitstorm...
Was thinking about this myself. As a libertarian I don't want regulation exactly but I do think we need a legal framework for handling how agents can act in the market. Like can an AI agent have "personhood" ? Or are the people they are agents for the responsible party? Or are the companies that developed the AI models? Who owns liability if an agent does something harmful?
These issues need to be nailed downbefore we have massive legal messes that prompt called for mu h more onerous regulaion
great questions and absolutely agree...these are institutional nightmares just waiting to happen whichever way the pendulum swings.
Is that the sound of the French government fleeing south and the thud of Wehrmacht boots tramping through the defenseless streets of Paris?
Like, c'mon! How hard is it to say, "Any paper that scores 50% AI or more on Pangram is an automatic fail and any paper in the 25-49% AI range will be subject to additional critique?" Pangram may not be accurate at a sentence by sentence level, but it is extremely accurate at judging large blocks of text (700 words or more). Correct me if I'm wrong, but I thought it was the business of educators to teach students how to think for themselves. That includes writing and the excruciating articulation of ideas which is really a student's first test of them, manifesting them into sentences and connecting concepts with evidence in a step-by-step process. But this "AI encouragement" strikes me more as an abandonment of that process, much as the French government abandoned the people of Paris instead of fighting block by block. A student's capacity for critical thinking deserves our attention and our defense! But now it seems we are granting permission to students to outsource their primary job to an unintelligent, probability calculator... unless, of course, we have only every cynically defined that job as submitting a certain quantity of assignments and checking a certain order of boxes on the way to receiving a perfunctory participation grade.
It used to be that plagiarism was the thing that plagued professors. Now it's AI. Different foe, but same principle. And much like plagiarism, after a while, you don't need even need Pangram to discern a lot of AI writing. I'm glad my professors fought for academic standards that help students reach and grow. It would be a shame if the evolution of technology causes the next generation of profs to throw up their hands in futility just because they don't want to have a couple of difficult conversations.
Pragmatism is a virtue, up to a point. But let's not delude ourselves for what this is, wholesale surrender thinly veiled as fairness.
The requirement of at least one responsible human in the loop should be part of every automated system. The old "computer error" excuse should never have been tolerated.