Students Around the World are Using AI-Powered Smart Glasses to Cheat on Tests 52
Students are using AI-powered smart glasses to cheat on tests, reports CNN. "And in East Asia's test-obsessed societies, where a single exam could impact the trajectory of a student's future career and social status, educators are scrambling to get ahead of the problem."
Already, countries are stepping up inspections for test-takers. For China's grueling annual college entrance exam earlier this month — which more than 10 million hopefuls take each year — authorities required screening of all glasses. In the United Kingdom, the head of England's exam watchdog warned earlier this month that AI glasses and smart devices like earpieces could worsen cheating in exams... [T]wo incidents in South Korea were the country's first reported cases of cheating with AI glasses... In Taiwan, the university where a prospective student was caught cheating is now reviewing rules and standard operating procedures for AI eyewears during examinations.
But experts worry these individual cases point to a more widespread issue. "If we're seeing a few cases being reported, we're seeing a lot more cases not being reported," said Thomas Corbin, lecturer at Deakin University in Australia, who has conducted research around the usage of AI-powered glasses and other smart devices in academic assessment. With the rapid development of AI technology, however, smart glasses are becoming slimmer, less noticeable, while integrating AI models that can operate independently with connectivity, raising concerns not only about exam integrity, but also about broader privacy risks... "Wearable AI is as much of a challenge to exams as ChatGPT was to essays in 2022 and I just don't think there is any real way that we can reliably have exam practices moving forward," Corbin said.
But experts worry these individual cases point to a more widespread issue. "If we're seeing a few cases being reported, we're seeing a lot more cases not being reported," said Thomas Corbin, lecturer at Deakin University in Australia, who has conducted research around the usage of AI-powered glasses and other smart devices in academic assessment. With the rapid development of AI technology, however, smart glasses are becoming slimmer, less noticeable, while integrating AI models that can operate independently with connectivity, raising concerns not only about exam integrity, but also about broader privacy risks... "Wearable AI is as much of a challenge to exams as ChatGPT was to essays in 2022 and I just don't think there is any real way that we can reliably have exam practices moving forward," Corbin said.
They are only cheating themselves (Score:5, Insightful)
The other take-away is that tests are pretty useless anyways. I have known that since I started teaching.
Re: (Score:3)
I would have loved to hear "tests are pretty useless" 40 years ago -- I was terrified of tests, and scraped through university.
I got 10% on one mid-term, which earned me an interview with the VP Academic of the university, who was also the prof. He grilled me on the material (thermodynamics), and I did just fine. He said, "You seem to know the material -- do you just not do well on tests?", and I said, Nope, never have. His answer: "Huh."
I had to continue to write exams, and continued to do poorly. I would
Re: (Score:2)
Your experience does not surprise me. While I test well (I research the examiner, past exams and their "way of thinking"), I think test-preparations are basically a waste of time and rarely serve as additional repetition. What I made good experiences with is student presentations, in-semester projects and oral exams.
But here is the dirty secret: Written exams are a lot less work and you can even fob off a lot of the work to assistants. There are also "test banks" (tests sold confidentially with books) and t
Re: (Score:2)
I would have loved to hear "tests are pretty useless" 40 years ago -- I was terrified of tests, and scraped through university.
I got 10% on one mid-term, which earned me an interview with the VP Academic of the university, who was also the prof. He grilled me on the material (thermodynamics), and I did just fine. He said, "You seem to know the material -- do you just not do well on tests?", and I said, Nope, never have. His answer: "Huh."
I had to continue to write exams, and continued to do poorly. I would have been thrilled to have a one-on-one session instead of an exam.
Yea. A number of my engineering profs offered an oral exam vs a test if you wanted it. Later in grad school, they didn't and one class (accounting) I did poorly on tests (I'm an engineer, close is good enough for airplanes) and would go over tehste with my prof, who said I know the material I just have trouble adding and subtracting...
Re: (Score:2)
At which point you should be slapped and your engineering creds revoked. Because that's where you break out your HP (calculator) and do all the addition and subtraction (using RPN, you are in engineering, right?) on it.
Everyone should be comfortable with a basic HP-12C. Fancy people have an HP 48, 49, 50, and even an HP Prime.
Re: (Score:2)
At which point you should be slapped and your engineering creds revoked. Because that's where you break out your HP (calculator) and do all the addition and subtraction (using RPN, you are in engineering, right?) on it.
As I said, I'm an aeronautical engineer. Close is good enough, 6 is close enough to 4 for large values of 4, and in case, it'll still fly... If it doesn't just add bigger engines. As I told my accounting prof, I'll just hire an accountant (and I did).
Everyone should be comfortable with a basic HP-12C. Fancy people have an HP 48, 49, 50, and even an HP Prime.
Mine is an HP45. Later got an 41CV. RPN means never having family borrow your toys...
Re: (Score:2)
Re: (Score:2)
Oral exams are the best way, but how do you do them when you have hundreds of students? That's the problem.
Re: (Score:3)
Oral exams hit their own problems of tester bias. In engineering, women tend to do worse on oral exams than written. That had been chalked up to just differences in learning styles, but I heard that an investigation at my university found it was the way instructors phrased questions or pressured for answers that changed. So you need someone giving the oral exams who is trained to ask the questions the same way -- except that standardization is exactly what makes oral exams better... you can probe for what t
Written tests are better (Score:2)
There has to be a measurement of area of study competency and knowledge in the college process.
Written tests are better as they reduce the bias for or against the test taker due to attractiveness, manner of speech, gender or background.
Oral tests fail in that they allow the charming and attractive persons and advantage over others as well as introducing a bias from the test grader.
AI graded test will be worse and less reliable due to the bias of grading for the average opinion, though, fact, or repeated pat
Re: (Score:2)
Aren't they all about prepping the student for the world and their future in it?
If you're someone who fails at written tests, hope you at least learned that much and go into a field where you don't need to do work similar to that. Ex. don't be a supervisor that has to fill out a lot of paperwork. But if you excel at test taking and book learning, maybe look into something fitting (legal?).
OTOH, if you're someone who doesn't do well with oral exams, learn that and don't go into such jobs. Ex. don't be a sale
Re: (Score:2)
There is a problem (I wouldn't imply gender bias) with some people not being good at oral exams, but still a good examiner can notice that and help them. Sometimes they can change the topic and ask about another part of the curriculum, or give a hint, or a partial solution to assess if the person has the knowledge and is only stumbling.
Also written exams are comparable and can be challenged better later on, so you avoid problems with the examiner being an asshole. But with good teachers the average grade in
Re: (Score:2)
Talking about STEM here, where you can take notes "Topic X: knew Y" (which is what the co-examiner does) and have little room for interpretation. This may be more complicates in areas without clear example solution.
Re: They are only cheating themselves (Score:2)
If it were that easy, companies would have far fewer problems with job interviews.
Re: (Score:2)
Re:They are only cheating themselves (Score:5, Insightful)
What's the alternative? Completely subjective grades that are assigned to the students by their teachers?
(That was meant to be rhetorical, since that is obviously even more worthless).
Given the economic opportunities that grades open up, I don't think it is fair to say "they are only cheating themselves." They are cheating others out of work and/or scholarship money, too.
Re: (Score:3)
Here is another well-kept secret: Grades _are_ always subjective. By unfair (but legally sound) correction, you can massively change grades. I have experimentally done it (not for actual grades but with actual exams) just to see how large my responsibility to be fair really is. It is huge. I now mix exams after correcting each question and I make sure not to look at the names, just to remain impartial and fair.
Incidentally, the same is possible for automatic grading. How? You have a look at the exercises. T
Re: (Score:2)
Well, if we realize that the North American system of education is rather unusual in that your future is really put in the hands of the student, and go with the system used around the world where they have a major test at
Re: (Score:2)
It's not a bad idea to put some pressure on kids, even if they crack. In engineering, it's important to test materials until breaking point. In human occupations where stress is an important day to day component (eg medical ER or brain surgery) it's best to see the kids fail *before* they enter university, spend 7 years studying, and find that they can't cope with the pressure after they become interns.
People need to know themselves before they can pick a path, and society needs to know they can handle t
Re: (Score:2)
Incidentally, the same is possible for automatic grading. How? You have a look at the exercises. Then you can design questions so that some specific participants fail and others pass. This can also be done without really realizing it.
True story - my grade 11 chemistry teacher specifically set up the pop quizes to spank otherwise high-performing students. He was extremely up-front about it, and each of us only got dinged once (and it's on a smaller quiz so it didn't materially affect grades). His point was that none of us were perfect, and a little humility was a good thing. I still remember taking the quiz, handing it in with the comment "was my turn, wasn't it" and his shit-eating grin.
Damn, I wish my kid had teachers like him.
Nostalg
Re: (Score:2)
Oh it's "there's no alternative" now? Why this preoccupation with testing? Where do you even get your ideas from? You can't conceive of anything else?
Teaching isn't the same as testing. How about you start with a conversation with the student, like "what did you find simple, and what did you find difficult studying the material?" and go from there.
Re: (Score:2)
That would require, you know, actual insight and maybe even experience! That is soooo 1990. Today we just make statements of fact with no clue what the facts are and how things actually work.
Re: (Score:2)
Oh it's "there's no alternative" now?
No, there was always no alternative. Testing is the only way to quickly at a point in time assess knowledge. Incidentally testing is not done in trade apprenticships, but I don't think most employers are willing to throw 3 years into a university educated graduate as a trial to see if they are any good.
Re: (Score:2)
Testing is the only way to quickly at a point in time assess knowledge.
Except it doesn't usually do that. Most testing quickly at a point in time assesses your ability to pass a test. It doesn't test your ability to apply the skills and knowledge in a real-world scenario, because the tests are so different from the actual work. For the few tests which are truly applied, yes, they do that. For the rest, no, too much of it is biased towards exam-taking.
Re: (Score:1)
Re: (Score:2)
I used to allow pocket calculators. You know what? They were a _detriment_ and made people slower. Sure, I only have entropy estimations and size estimations in there and I do not require an exact answer, but after I saw CS students struggle to do a binary log on a regular calculator, I have stopped allowing them.
Re: They are only cheating themselves (Score:2)
Re: (Score:2)
Yes. I am fortunate that I am doing academic teaching and doing in a dynamic field (IT security) so I can compensate for some things and select myself what I teach. But when I do a reasonable course, certainly not a great one and students tell me "most interesting course of the year", it becomes very clear that something is deeply wrong.
Re: (Score:2)
I'm also a teacher...and how do you figure? They're imperfect, there's issues around them...but useless? Ultimately, if you create a test that's even semi-decent, students who know the material well will get an "A," students who have no fucking clue what's going on in the class will get an F, and other students will get somewhere in between. Isn't that what you'd want?
Re: (Score:2)
What relevance does "knowledge" have today? Right, not a lot. But what about skills? Pretty critical, IMO. Yet you test for knowledge, not for skills. Hence "useless". This is obviously simplified but covers the gist. If you teach, you should _really_ understand that.
Re: (Score:2)
Sounds like you should learn how to make better tests.
Really? (Score:3)
Really? Where are the proctors?
There are no smart glasses on the market that can hide what they are. Everyone and their dog can identify smart glasses from across the room. So, how is this a thing?
Furthermore, if we are supposed to believe and accept that they are going unnoticed and getting in to exam rooms, how are we able to get these numbers on cheaters undetected?
This is Facebook drivel.
Re: (Score:2)
Furthermore, if we are supposed to believe and accept that they are going unnoticed and getting in to exam rooms, how are we able to get these numbers on cheaters undetected?
This argument has no merit. Provide credible anonymity and the right environment, and people will admit it. Works a bit like exit-polls.
Re: (Score:2)
Do we have any of those things? No. All we have is your supposition based on an, admittedly plausible, scenario of your own contrivance.
Yet you have the audacity to say my statement has no merit. LOL!
Re: (Score:2)
However, if we commit the ultimate Slashdot sin and actually read the article, we find out they never claimed to have numbers. The closest they come is
which is pretty clearly and openly straight speculation. It is y
Re: (Score:2)
Did I claim they were doing this? Nope. That is all your imagination. I just pointed out that, contrary to your faulty reasoning, this can be done.
"Audacity" indeed...
Re: (Score:2)
Agreed. (Score:2)
Testing can be done under conditions that make cheating impractically hard. It just takes a little effort on the part of those giving the tests.
Block the radio waves (Score:2)
For now at least, wearables require radio frequencies to work.
So block the radiowaves.
https://www.homedepot.com/pep/... [homedepot.com]
Paint the room with EMF shielding and you do not have to worry about any connections - not cellular, not bluetooth, nothing.
Re: (Score:2)
That's why they talk about AI. This can (or will be able to soon) run on the device itself. On the other hand, one might also be able to just access a textbook on the device itself with the same result.
No one saw that coming (Score:2)
At least not if they were wearing AI glasses imposing VR advertising for next month's improved AI glasses.
Obligatory joke for low hanging fruit.
Adam Smith's biggest mistake was stealing the invisible hand's cloak of invisibility. Mostly downhill since then, now at AI-enhanced speeds.
They are only cheating themselves (Score:2)
Kids have it easy today (Score:2)
more risk in exams, less broader risks (Score:2)
The conclusion seems slightly off. Using local AI instead of cloud services increases the risk of cheating in exams (if they try to shield rooms against network access), but at the same time increases privacy for everyone who is filmed by the classes.
200 year old tech (Score:2)
Put a Faraday cage around the classroom. Hell, put one around the whole building. Hell again, put one around the whole campus, and allow communications through dedicated gateways only. (We called them "internet proxies" in the stone ages.)"
All Signs Point To (Score:2)
Re: (Score:2)