Slash Boxes

SoylentNews is people

posted by hubie on Monday March 11, @03:43PM   Printer-friendly
from the complaints-department-5000-miles-> dept.

In a notable shift toward sanctioned use of AI in schools, some educators in grades 3–12 are now using a ChatGPT-powered grading tool called Writable, reports Axios. The tool, acquired last summer by Houghton Mifflin Harcourt, is designed to streamline the grading process, potentially offering time-saving benefits for teachers. But is it a good idea to outsource critical feedback to a machine?
"Make feedback more actionable with AI suggestions delivered to teachers as the writing happens," Writable promises on its AI website. "Target specific areas for improvement with powerful, rubric-aligned comments, and save grading time with AI-generated draft scores." The service also provides AI-written writing-prompt suggestions: "Input any topic and instantly receive unique prompts that engage students and are tailored to your classroom needs."
The reliance on AI for grading will likely have drawbacks. Automated grading might encourage some educators to take shortcuts, diminishing the value of personalized feedback. Over time, the augmentation from AI may allow teachers to be less familiar with the material they are teaching. The use of cloud-based AI tools may have privacy implications for teachers and students. Also, ChatGPT isn't a perfect analyst. It can get things wrong and potentially confabulate (make up) false information, possibly misinterpret a student's work, or provide erroneous information in lesson plans.
there's a divide among parents regarding the use of AI in evaluating students' academic performance. A recent poll of parents revealed mixed opinions, with nearly half of the respondents open to the idea of AI-assisted grading.

As the generative AI craze permeates every space, it's no surprise that Writable isn't the only AI-powered grading tool on the market. Others include Crowdmark, Gradescope, and EssayGrader. McGraw Hill is reportedly developing similar technology aimed at enhancing teacher assessment and feedback.

Related stories on SoylentNews:
SWOT Analysis of ChatGPT in Computer Science Education - 20240215
OpenAI Admits That AI Writing Detectors Don't Work - 20230911
An Iowa School District is Using ChatGPT to Decide Which Books to Ban - 20230817
A Jargon-Free Explanation of How AI Large Language Models Work - 20230805
Why AI detectors think the US Constitution was written by AI - 20230718
Dishonor Code: What Happens When Cheating Becomes the Norm? - 20230301
Amid ChatGPT Outcry, Some Teachers are Inviting AI to Class - 20230221
Seattle Public Schools Bans ChatGPT; District 'Requires Original Thought and Work From Students' - 20230119
ChatGPT Arrives in the Academic World - 20221219

Original Submission

This discussion was created by hubie (1068) for logged-in users only, but now has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 5, Insightful) by Thexalon on Monday March 11, @06:10PM (16 children)

    by Thexalon (636) on Monday March 11, @06:10PM (#1348278)

    Not just this specific instance of it. Like, the whole concept of grading sucks. It's a poor attempt at applying an industrial model to an activity which is fundamentally not like factories in the slightest. There have been studies that strongly suggest that introducing grading is the fastest way to destroy a child's interest in learning things just for the sake of knowing those things.

    And no, I don't think standardized tests or IQ get to the crux of the problem either: It turns out that reducing human brains to an INT stat just doesn't capture how smart a kid is or what specifically they're capable of, but it does do an excellent job of capturing how much money and schooling the kid's parents have.

    The only thing that stops a bad guy with a compiler is a good guy with a compiler.
    Starting Score:    1  point
    Moderation   +4  
       Insightful=2, Interesting=1, Underrated=1, Total=4
    Extra 'Insightful' Modifier   0  
    Karma-Bonus Modifier   +1  

    Total Score:   5  
  • (Score: 3, Interesting) by Mykl on Monday March 11, @10:04PM (4 children)

    by Mykl (1112) on Monday March 11, @10:04PM (#1348323)

    I don't think it's a universal truth that kids are discouraged from learning by being scored.

    Anectodally, one of my kids was content to drift along for the early years of his schooling, and would only learn what he had to in order to have enough marks to stay out of trouble with us. He discovered his passion for learning much later (mid-teens), but would not have been able to do so if we had just left him to his own devices. He is now an eager student who is highly engaged with his schooling.

    Also anecdotally, another of our kids has a friend whose parents see school as 'optional'. He is allowed to attend if he wants to, and basically chooses not to unless there is a special event or activity on that day. He is a lovely kid, but is functionally illiterate and obese (he is not encouraged to exercise or get outside when he stays at home). I am convinced that this kid would be in a much better position to succeed in later life it he had been pushed forward (grading etc) rather than just being allowed to learn at his own pace.

    I agree with you that there are lots of different ways in which people can be smart, but I disagree that we should do away with grading. How else can one determine whether they have properly learned a particular skill, and given feedback on how to improve?

    • (Score: 2) by Thexalon on Tuesday March 12, @02:36AM (3 children)

      by Thexalon (636) on Tuesday March 12, @02:36AM (#1348354)

      It's not universal, but there are studies [] out there suggesting that testing is a demotivator. What's certainly true is that testing means that the kid is learning whatever the test says is important, rather than what might actually be critical or exciting for that student.

      The only thing that stops a bad guy with a compiler is a good guy with a compiler.
      • (Score: 2) by Mykl on Tuesday March 12, @02:55AM (1 child)

        by Mykl (1112) on Tuesday March 12, @02:55AM (#1348356)

        Let's say that we want to do away with testing as it de-motivates. How else do we determine whether a child has learned what they need to know? Any sort of exercise that checks that they have learned something is in effect a test.

        I don't think you are saying that we should abandon all checks to confirm that kids know what they need to know. But what is the alternative?

        Specific scenarios: Qualifying as an Engineer. Being a surf lifesaver. Obtaining a drivers license. Earning an academic scholarship to a prestigious school

        • (Score: 2) by Thexalon on Tuesday March 12, @11:57AM

          by Thexalon (636) on Tuesday March 12, @11:57AM (#1348392)

          There was a time not that long ago when all of this stuff was extremely informal. As in, your teacher would evaluate your work as a whole and at some point tell you you were ready for the next class or next level of work. For university degrees, that would probably be a council of all the faculty in the subject you were studying who would decide to grant you a degree (or not). Letter grades and later numeric grading were experimented with in the 1800's and only became widely adopted in the early 1900's, and then as now the primary purpose was to figure out things like who got their tuition paid for and who had to chip in money that they might not have to get an education. So more-or-less coinciding with the Industrial Revolution, and more-or-less attempting to apply industrial factory thinking to the process of learning because standardization was considered the most important thing.

          Final exams were apparently an earlier phenomenon, and I don't think we should get away from, say, professional licensure exams that exist today. But I do think that learning is one of those areas where trying to reduce everything to conveniently spreadsheetable metrics is at best unhelpful and at worst counterproductive. The fact is that if you handed a good teacher a list of students in their class(es) and asked them to rank their understanding of the material from "grokked it" to "complete know-nothing", they all could do so without needing to look at test scores.

          The only thing that stops a bad guy with a compiler is a good guy with a compiler.
      • (Score: 3, Insightful) by anubi on Tuesday March 12, @02:57AM

        by anubi (2828) on Tuesday March 12, @02:57AM (#1348357) Journal

        The most destructive thing I have seen ranking/ grading people do is plant in them the idea that they are better than everyone else.

        Once they get that "superiority" meme planted in them, the "Stanford Prison Experiment" mentality comes into play, and the "teamwork" meme of the group rapidly disintegrates into a "dog-eat-dog" situation, everybody trying to compete to be the lead dog, everyone trying to delegate, no one's doing the work.

        I've seen this happen. I consider this relentless ranking one of the main tools used by management to absolutely ruin an organization.

        It turns the workplace back to grade-school where everyone hates the "teacher's pet".

        "Prove all things; hold fast that which is good." [KJV: I Thessalonians 5:21]
  • (Score: 4, Interesting) by aafcac on Monday March 11, @10:20PM (9 children)

    by aafcac (17646) on Monday March 11, @10:20PM (#1348328)

    I've got a Masters in Education Studies and teaching is rather complicated. Testing and evaluation are particularly important as you tend to get what you test. If you've got a really well-designed testing program for a class, you'll get students that are encouraged to put effort into learning. If you have a poorly designed one, like any ML based one is likely to be, then you get a lot of issues.

    Testing and evaluation needs to be something that is thought about near the beginning of the class design. Realistically, you're better off farming out the content exposure and teaching to a 3rd party than the evaluation, there are any number of credible sources of information out there and if students are exposed to multiple ones they will likely get a good grounding in the topic. But, if the evaluation has flaws in terms of rewarding lower levels of learning, or allows for the system to be gamed, then that's what you'll get.

    I do think that ML can still be of value, however, it needs to be used appropriately. Having the ML spit out a grade is asking for trouble, but having the ML spit out an evaluation and having the student explain why the ML's opinion is valid, or not, and give some supporting reasons is a valuable exercise in nearly all classes.

    • (Score: 3, Interesting) by Thexalon on Tuesday March 12, @03:17AM (8 children)

      by Thexalon (636) on Tuesday March 12, @03:17AM (#1348362)

      I've got a Masters in Education Studies and teaching is rather complicated.

      A bit about my background: One of my grandparents was an influential professor of education at Harvard, whose work on early childhood is still affecting how language is taught in the primary grades today (among their projects: Helping to develop what was then a new show for PBS combining entertaining puppets and educational content that you are probably familiar with). One of my parents taught high school for about half of their career.

      During my career, I've sometimes gone into tech education and worked with groups of kids learning the basics of programming.

      And at no point have any of them, nor any teacher I've ever chatted with, believed that testing provided any benefits at all to children. What I see as the main role of testing in our educational system today is helping the adults identify which kids we've decided it's OK to discriminate against. "You get A's, so you get to learn the skills that will get you towards university. You get C's, so we've decided there's no way you're going to be able to understand the stuff we're teaching our A students, so we're going to ship you off to a different building where you can learn to flip burgers." Oh, and by the way, it's nigh-impossible for the children of parents who went to university to end up in the burger-flipper track (their C's will lead to diagnosis with a learning disability that allows for accommodations that make things easier until they're getting A's again), and also fairly unlikely for the children of parents who were burger-flippers to end up on the university track.

      The only thing that stops a bad guy with a compiler is a good guy with a compiler.
      • (Score: 2) by Reziac on Wednesday March 13, @01:21AM (2 children)

        by Reziac (2489) on Wednesday March 13, @01:21AM (#1348484) Homepage

        "What I see as the main role of testing in our educational system today is helping the adults identify which kids we've decided it's OK to discriminate against."

        This was very definitely not how my many wonderful teachers thought. Rather, that they needed to work harder on that student, at least if the student expressed even the vaguest interest in learning. Or determine if maybe the student needed an entirely different class.

        But I was in school before teaching to the test.

        And there is no Alkibiades to come back and save us from ourselves.
        • (Score: 2) by Thexalon on Wednesday March 13, @02:56AM (1 child)

          by Thexalon (636) on Wednesday March 13, @02:56AM (#1348498)

          It also matters quite a bit where you were in school. For example, my school system was a wide mix of backgrounds, and a lot of effort was put into tracking kids into the "right" spots. And I couldn't help but notice a strong correlation between mom & dad's money and how smart the kids allegedly were.

          I've seen very different dynamics in play in, say, wealthy suburbs of Boston, and one of the rougher schools in inner-city Cleveland.

          I had some truly wonderful teachers too. But they were working in a system that did not reward them for helping the worst students.

          The only thing that stops a bad guy with a compiler is a good guy with a compiler.
          • (Score: 2) by Reziac on Wednesday March 13, @03:37AM

            by Reziac (2489) on Wednesday March 13, @03:37AM (#1348502) Homepage

            I went to school mostly in the Northern Wastes of Montana. We were very slow to get the rot, and quality teaching persisted well after the rot had hit more 'progressive' areas. (Whole Word Recognition hit Minnesota at least 15 years earlier.) Was still not seeing negatives here when I moved to SoCal in 1984... where it was immediately evident that schools were not all they could be; I was astonished by the vapid and emptyheaded kids. Now back in Northern Wastes... difference is again obvious.

            So, yeah, I'm sure it's variable everywhere.

            I'm glad to have had the educational experience I did.

            And there is no Alkibiades to come back and save us from ourselves.
      • (Score: 3, Interesting) by aafcac on Wednesday March 13, @12:35PM (4 children)

        by aafcac (17646) on Wednesday March 13, @12:35PM (#1348537)

        That just goes to show that in every field there are incompetents that don't know what they're talking about. Every model of learning that I've ever seen involves a point where there is evaluation, either internal or external. It's literally the difference between repeating the same stupid mistakes over and over versus improvement. It's also one of a short list of reasons that we have professional educators and schools dedicated to learning. Others being thoughtful selection of source material and the social learning that comes of spending time with other kids of a similar age. It's also been known for ages that you need to take a stab at the doing the thing before checking the answer. Without trying first, the proper neurons won't fire, and without verifying it afterwards you get random ML style learning that may or may not have any sort of meaning to it.

        The evaluation process addresses a massive issue with learning. Apart from relatively simple things where you can see whether or not it worked, people are terrible at self-evaluation. It's why basically everybody out there suffers from some combination of Dunning-Kruger or imposter syndrome. Knowing what precisely you know tends to require knowing more than you know in order to accurately evaluate it.

        The other bit about it, is that it's supposed to be one of the earliest considerations when teaching because if you're not integrating that at the beginning of the process. The testing and evaluation, when done right, is regular enough that students don't waste a lot of time guessing about why things aren't working. They can then focus on moving more in the direction of something that is known to work. If they aren't being given that, then there's absolutely not reason at all to be in school at all. If it's just being info dumped on and being given a random test so that they've been "tested" then there's no point in schooling at all.

        I'm curious precisely what those people think that learning is if it doesn't involve integrating testing and evaluation into the process so that students don't waste a bunch of time on things that are known not to work. OTOH, this sort of attitude does explain why I've wasted so much time over the years working with students that wouldn't have a problem if not for bad teaching. So much time undoing the damage before I even get to teaching anything new.

        • (Score: 2) by Reziac on Wednesday March 13, @02:20PM (1 child)

          by Reziac (2489) on Wednesday March 13, @02:20PM (#1348555) Homepage

          "It's also been known for ages that you need to take a stab at the doing the thing before checking the answer."

          Or why we learn more from our mistakes than from our successes?

          For some years I was the hardware bloke for a big user group. Got in a bunch of retired middle school PCs, still loaded with "educational" software. Being curious, I took a stab at it, and quickly realized the software did not teach the subject at all; rather, it taught how to get the software to spit up the desired response.

          And there is no Alkibiades to come back and save us from ourselves.
          • (Score: 2) by aafcac on Wednesday March 13, @08:28PM

            by aafcac (17646) on Wednesday March 13, @08:28PM (#1348623)

            That's definitely related. There's simply a lot more things that can be learned from failure than success to begin with. There may only be one way of succeeding at a task, or there may be a few. Edison failed to invent a durable light bulb in hundreds of different ways before he found one way that worked. These days, there are a few other designs that were developed later on and now we've mostly abandoned it entirely for LEDs.

            As far as software goes, educational software is usually a hunk of garbage for anything other than rote memorization of things like math facts. I fondly remember games like Number Munchers, Where in the World Is Carmen Sandiego and the Super Solver's series. They were relatively entertaining, but they were also fairly limited in terms of the content that the students would get out of it. It wouldn't take that long before it was just about the game and not about the learning. IMHO, the software is best for things where you need to encourage the students to practice repetitive skills a lot. I loved Mario Teaches Typing when I started to learn to type because it just made a bunch of sense. Learning to type is kind of tedious, and the game did a great job of scaffolding the process and making it engaging. By the end, you would be typing regular text rather than just hitting the right keys.

        • (Score: 2) by Thexalon on Wednesday March 13, @04:45PM (1 child)

          by Thexalon (636) on Wednesday March 13, @04:45PM (#1348585)

          I'm not suggesting "no evaluation". I'm suggesting that grades aren't a good form of evaluation. Compare and contrast these two forms of evaluation:

          "Billy shows a strong ability for writing poetry including intuiting advanced rhythmic concepts and slant rhymes, and is comfortable with arithmetic, but struggles with more abstract math like algebra, and also has a hard time with rote learning such as knowing the location of countries and the names of their capital cities."
          This is information you would want if you were teaching Billy with the goal of Billy both being able to excel in some things and have a good-enough understanding of everything else to at least manage the challenges life puts in front of him. For example, a geography teacher with that information would be better off trying to teach Billy about Ethiopia by giving Billy some basic information about the country and asking him to write a short poem or stanza about it and its capital, rather than having him draw a map of East Africa. Also notice that this kind of approach suggests that Billy's geography exercises should probably be different than his classmate Jane who would learn this better via the map. Also also notice that if you give Billy and Jane different assignments, they're in a position to cooperate with each other, because both Billy's poem and Jane's map will help in understanding Ethiopia.

          "Billy has a GPA of 2.97, and a class rank of 254 of 759. His test score went from 456 to 537 over the last 3 months."
          The second evaluation is about something very different, namely how we should set up Billy's education to maximize the gain in test performance while minimizing the cost in time and effort spent on Billy to achieve that gain. And indeed, they might adjust to not focus on Billy at all, but instead on Billy's classroom as a whole, because the goal is to maximize the mean test performance, and maybe it's better to neglect Billy, because Jane can provide a much better test score increase at lower costs and thus make the teacher and school look better to the administration system. And of course it has ranked Billy and Jane to decide once and for all which of them has the superior brain and thus which should get the awards and rewards associated with that. This process, properly applied, will help in refining the assembly line to produce more qualified workers and students for the next level of education. You see what I mean about industrial thinking I hope.

          The only thing that stops a bad guy with a compiler is a good guy with a compiler.
          • (Score: 2) by aafcac on Wednesday March 13, @08:18PM

            by aafcac (17646) on Wednesday March 13, @08:18PM (#1348622)

            What does this have to do with any of what I posted? At no point did I specify that there needed to be letter or number grades, nor did I say that people need to be ranked as part of the process.

            Testing and evaluation is a rather broad category that includes everything from one ranked test with a letter grade to informal measures of engagement like student eye contact and body language. Simply counting the number of times students offer to answer questions in class or ask questions would qualify as well.

            The specific type, frequency and weighting of the evaluation should depend a lot on the subject and the ultimate goals of the coursework.

  • (Score: 2) by Reziac on Tuesday March 12, @03:20AM

    by Reziac (2489) on Tuesday March 12, @03:20AM (#1348363) Homepage

    When I was in school (back before electricity) our test scores were often posted for everyone to see.

    One of our primary motivators was not being seen with a score below our known capacity.

    How is this different from any other sort of scoring?

    And there is no Alkibiades to come back and save us from ourselves.