Gizmodo: It’s Breathtaking How Fast AI Is Screwing Up the Education System

[Fitz]

Career teacher here.

I think it's worth noting that AI detection programs are notoriously unreliable. They return such a large number of false positives that holding results against students is untenable.

[Renegade7]

Career teacher here.

I think it's worth noting that AI detection programs are notoriously unreliable. They return such a large number of false positives that holding results against students is untenable.

Either they know what the wrote or they don't in my experience.

You can only bs so much what you said you cited.

With turnitin it could detect AI. But it was rare I saw something that was actually flagged. The detection around citations was on point imo.

This is almost 2 years ago...that's a lifetime with respect go how fast AI is moving.

There has to be Governance around this and clear guidance in how to go about it...I think you used AI because a tool told me is different then huge amount of flagging and asking questions to prove they did or didn't write it themselves.

At the college I was teaching at you had to make a case that someone was plagiarizing, you couldn't jus witch hunt your students.

[mcsluggo]

this same discussion/debate happens everytime there is an information advance.

nobody will do research because encyclopedias now summarize everything for you
nobody will read anympre because you can watch the movie version/read the cliff notes
nobody will understand math because the calculator replaced the abacus
nobody will understand math because scientific calculators replaced calculators
nobody will understand math because statistical programs replaced using the t-stat tables at the back of stats books
everything is cheating because they can find it using google

ai is just a new research tool. old profs will be stymied because they aren't used to it. for the next generation the skill will be how to ask the correct AI questions, how to parse the information, and then how to use follow up questions to head down the trail to get to what you actually need to know with minimal errors. it is just a new, more powerful tool. smart people will maximize the tool. stupid people will still be stupid but with more "information" (and misinformation) squandered at their disposal.

[The Sisko]

When I was teaching at my Alma mater we had a tool called TurnItIn that we could put any paper into and tell the ratio written by AI and even if they used the sources they said they did.

https://www.turnitin.com/

This should be mandatory in every education system now.

I just finished my MPA last fall. I used AI extensively to outline papers for me because ADD, but I also let it do probably 65ish% of the actual writing. However, after the AI had done its thing, I had to go through it to root out any nonsense it may have thrown in, then some select rewriting to make it my own plus a few edits to add things specifically designed to throw off tools like Turnitin. I’d then follow that up by running it through Turnitin and another AI checker to be sure.

For me it was really about shortening the time to do the actual writing. For my capstone paper, I used the same strategy for the background info, but synthesizing it all into something to make my case about a topic I was/am passionate about was where the effort came in. That and making sure the sources I was citing were un/less biased and weren’t part of the “adoption abolition” crowd.

The thinking and analysis part is the real value add in the education experience IMO. If students approach it that way, I really don’t see the harm in using AI. When they enter a general bs question and hand in the output as-is is where I have a problem. Aside from academic dishonesty, one just has to be incredibly stupid AND lazy to try that.

[Renegade7]

@The Sisko

I can dig it.

A lot of people are stupid and lazy, I guess...

Too many are letting the AI lead the way and do it for them and not take anything away from it learning wise.

AI by itself isn't the problem right now, we are in how we use it versus how we try to augment what we doing for say efficiency sake, similar to what you mentioned.

I've seen in training that turnitin can catch someone translating a paper into another language then converting it back to English.

I'm not going to pretend I don't use chatgpt at work to write scripts, splunk queries, and better understand risk associated with different pieces of software in our environment, we all have to be cateful what we ask because hovernance is so far behind right now. AI is an incredible tool when used as a tool, not letting it do everything for you.

To keep from leaving this open ended, AI isn't going anywhere, so the need to catch unethical and rule breaking behavior has make every effort to keep up. People need to be afraid to try using AI to jus do their papers for them, this is a problem right now in large part because too many believe they can get away with and frankly they are, imo.

We have to stay honest and not pretend that even before generative AI was everywhere that everyone that cheated was caught.

I know what I saw when I was college, my Wife knew a third of the group of international students she was doing her masters with got expelled at the same time for getting caught cheating together, like 3-4 of them just like that. Things I saw in my undergrad still make me sick to this day. It hurts our industry (cybersecurity) more then it hurts them individually, big picture here.