OT College Student catches Prof using AI

https://currently.att.yahoo.com/news/northeastern-college-student-demanded-her-112924481.html

1 Like

I shall send this to my daughter. It was a total surprise to me…and a source of frustration for her…that the episode involving a domestic violence evaluation in her upcoming divorce was totally upended by her ex’s use of a generative AI programme. For anyone who didn’t catch my post a few months back, my hawk eyed reading of his online answers to a questionnaire (creating almost a whole life story that put him as the victim) realised something was up with “his” liberal use of the Oxford comma and gender neutral pronouns to refer to my granddaughter. Grammarly’s plagiarism detector threw up >90% plagiarised.

No consequences for this attempt to subvert the course of justice…yet. Maybe we need to look at it again!

1 Like

Just for the heck of it, I goggled “when AI creates a false narrative”. The page opened with the insight that there was no overview for the topic, but among the pages that did come up, there was this as an opener …

and this

…so, it’s not like there’s a big surprise about it all (as surprising as I found it)

1 Like

I’m not ready to call this intelligence: scrape the internet (or other content) and repeat it back to me in complete sentences.

The complete sentences give the appearance of intelligence, the actual content, not so much.

I am ready to call it artificial.

6 Likes

This should not come as a surprise. All the current version of AI does is statistical pattern matching without verifying the validity of the output. A Richard Feynman lecture explains, I paraphrase:

We start with a guess, an informed guess, but a guess non the less.

That’s where the Scientific Method kicks in, the hypothesis, the theory, the conjecture, the guess, is put to the test. Albert Einstein postulated that mass bends light. This was confirmed by experiment…

The Scientific Method is what AI is still missing.

The Captain

5 Likes

I’ve oftentimes wondered at the synchronicity/coincidence/Life imitates TMF with topics on these boards and this thread is no exception. It seems my clumsy “down the AI rabbit hole” attempts might have a future use…unfortunately.

Even though the daughter’s divorce is a past issue, the settlement…of marital assets, business assets, embezzlement and other criminal activity, and custody of my granddaughter…is not. Very not. Given that so much of a factual nature has been manipulated in a way that has made us question our perception of reality over the past year and longer, I’m starting to wonder just how much has been subjected to AI … and, more importantly, how would one know.

As suggested, AI is not actually very scientific a way to evaluate evidence. Even scientific subjects require a bit of caution…but what about disciplines that lack the testing strategies of the Scientific Method like, say, the Law? In an adversarial system, it seems to me that it’s the better cherry picking that wins the argument rather than the reliably reproducible evidence. I’m wondering and following where Google … and its possibly totally AI generated articles … sends me but it seems like there’s no easy way to detect any monkey business without specific knowledge??

Musing on this board for my own benefit (say with any suggestions for refining my search)…and maybe for others. I’m pretty surethat when I was musing similarly about personal experience with my delayed diagnosis of matters cardiovascular, it was seen as a one off anomaly. If GenAI really is as prevalent as it appears, any one of us could find ourselves on the losing side in a serious argument we never expected.

@Leap1 …this thread starter opened a can of worms.

3 Likes

Good question! Human Intelligence (HI) sorts subjects by domain and each domain has its own rules. The Scientific Method is just one domain and it might be subdivided by subdomains. Law might be subdivided by country or by religion. AI is in infancy.

Pattern matching is just raw input.

The Captain

2 Likes

My problem with this issue that’s irking me is that I can’t put a question together that gives me an answer to what I’m asking…or think I’m asking. Plenty of opinions and observations on use and merits/ethics of AI in information gathering and sorting. I guess the lecturer in the OP was using the technology for that. My question would be more pertinent to the student using the same technology to answer questions on an important test without revealing that upfront.

That’s what I was concerned with last November…when it came to light that the albatross had been determined to be the one who was the victim of domestic abuse and had been throughout the marriage!!! Based upon answers provided to an online questionnaire. The report was dismissed … but with no repercussions from such a blatant manipulation of the truth…and it only came to light because of the incongruity of the answers he provided and the **$50 words he used. Allowed to stand, my daughter could’ve lost custody of the granddaughter. It can’t be a first or only example and I would hope there are more reliable detection tools than a mother in law who can’t sleep and is willing to plough through a pile of unpleasant to read details.

** Those $50 words…make that ***$500 or so. In the kerfuffle after the husband had run the whole document through his plagiarism detector app, the daughter’s divorce doula asked me to compile a list of half a dozen of the fancy words contained in his answers that he’d be unlikely to use/understand…I guess to ask during the pre trial deposition. Never got round to asking when the time came but such mellifluous prose as ubiquitous, consequential and one that I had to double check…felicitating (meaning celebrating…not facilitating)

I’m still wondering how and what programme he must’ve used.

*** I haven’t put my bill in yet!

1 Like

So, @Leap1 , as chortle-worthy as your opening link is, I think I’ve got it matched…

Although dh gave me a demo yesterday with one of his text manipulation programs on the nuts and bolts of presenting simple statements "in the style of "… explaining to a 5 year old/preparation for scientific journal/as a narcissist/as a victim of narcissistic abuse (the plague blossom’s MO on this evaluation questionairre… Srsly!!) and I can now see how it’s doable…I’m still on the hunt. Specifically for examples of testimony creating an alternative life for a wannabe victim. Not there yet, but I found the above

2 Likes

Another little piece in the “Who thought THAT would be a good idea?”. I hope board members are blissfully unaware of the self perpetuating industry surrounding divorce/family law…particularly child custody. All manner of offshoots dipping their bread in the gravy boat. One such being communication apps for parents in acrimonious relationships to communicate on matters surrounding their children like dr’s appointments etc… Superficially a good idea and courts recommend them as they’re open to scrutiny by lawyers, therapists, and random hangers-on. The assumption being that if the “other side” can see how you communicate with your spouse, then folk are going to be less prone to outbursts and try to work in the best interests of a child.

I’ve seen how the Canker Sore has tended to try baiting the daughter on this venue (every so often successfully…like when he posted a series of very abusive messages/as if he were in the throws of a psychiatric breakdown and then claimed hed been hacked and Geri’s response was typical of her overreaction) but she’s become wise to his tactics…and likely use of ChatGPT (given the Strunk and White level of writing style … except when he’s pretending to be crazy, which itself is the act of someone who is, if you ask me!)

Well, I did wonder about how prevalent misuse by AI might be and I see that, if it’s not very at the moment, it soon will be…and less easy to detect than unlikely Oxford comma usage!

https://www.accessnewswire.com/newsroom/en/computers-technology-and-internet/ourfamilywizard-launches-new-ai-tool-to-support-healthier-co-pare-1025752

Oh my…I am being stalked!! I don’t mind one bit. I typed “AI and manipulation on Our Family Wizard” into Google just now…and a whole screed answering my question came up (probably AI generated, so…) I can’t fathom quite how to copy and paste the link but took a series of screenshots just in case the info changes to something I don’t want to read

A bit of a clunky way of sharing information but the weird thing is this was definitely not there on Google 2 days ago…else I would’ve saved myself a bit of time and a lot of extra information to clog my brain.

Now I’ve got this memorialized as is, I’m going to check back in a day or so to see how this might’ve changed.

1 Like