I've seen a few more in the wild too: GPTzero, ZeroCheating, and I myself have been working on one that tries to provide the same type of proof of effort but we're leaning more towards writing analytics and information about writing quality at CursiveTechnology.com.
There's a tension between transparency and privacy that's been playing out over the last 2 decades (especially in online higher ed), if we can use these to solve the conflict about AI or no AI while retaining writing as a valuable assessment option (and verifiable skill)
It'd been fascinating to research, develop, and implement. FWIW Typing biometrics has been around for 4+ decades for identity management/authentication, the accuracy is well documented.
It's use for authorship is more personal: as the writer you create it naturally every time and exert control over whether/if it works.
Grammarly's tool to be a positioned as a "personal integrity approach" (as a student I take responsibility to protect myself and actively commit to showing integrity). I'd say I'm long on personal/individual integrity and accountability (esp. if the same data can help me be a better writer).
Great breakdown and explanation Anna! It's really helpful to have this to learn more about these tools!
I do keep thinking about the watching the writing process happen. It's one of those things where with some folks, it's incredibly helpful to see and support. And also, in the wrong frame of mind, becomes harmful. And who gets to decide that and how is the thing that makes it challenging because it still often feels institutions and faculty by proxy get the power in that decision. It can feel too close to the "well, if you have nothing to hide, you should be fine with letting us search your car."
I have to wonder if Grammarly's next step is to create/step into the "plagiarism" market and be a competitor to TurnItIn...speculation--sure, but that projection certainly makes sense. If they make this work well for users (and normalize it as a practice), then sell the backend access to institutions, it would make $$$.
Thanks for the comment--I'm really happy the post seems helpful. Yeah, I appreciate your points and the comparison to police searches.
After some similar concerned comments on X, I am thinking some more about offering an alternative to process tracking: if they prefer not to share document history they could meet with me (or perhaps record a video) to chat about their writing process. Of course they might feel shy about doing that, and it's extra work, but I'm not sure what else could be an alternative...
I think Grammarly already offers plagiarism and AI detection. The profit incentives are definitely key to watch. I don't know much about Grammarly for Edu. I haven't worked anywhere that would consider paying for it, I don't think...
yeah--I think these tools definitely open up the exploration of process and evaluating processes to understand/appreciate outcome.
It makes me wonder what there might be to learn from other disciplines about how we do this (Math and "show your work" or Art--which of course, they also face that problem of capturing the process well)...
Mmm, great point. In writing instruction we do often assign reflections on the writing process, and that in itself is one of the strategies that helps some with deterring misuse. I'm still looking for the best ways to link this kind of process tracking with process reflection. It seems like it should be a native part of the process tracking report app that it nudges the author to reflect and annotate.
Great summary!
I've seen a few more in the wild too: GPTzero, ZeroCheating, and I myself have been working on one that tries to provide the same type of proof of effort but we're leaning more towards writing analytics and information about writing quality at CursiveTechnology.com.
There's a tension between transparency and privacy that's been playing out over the last 2 decades (especially in online higher ed), if we can use these to solve the conflict about AI or no AI while retaining writing as a valuable assessment option (and verifiable skill)
Interesting! I would worry about writing analytics to identify authorship in terms of accuracy and false positives, but I haven't looked at this yet.
It'd been fascinating to research, develop, and implement. FWIW Typing biometrics has been around for 4+ decades for identity management/authentication, the accuracy is well documented.
It's use for authorship is more personal: as the writer you create it naturally every time and exert control over whether/if it works.
Grammarly's tool to be a positioned as a "personal integrity approach" (as a student I take responsibility to protect myself and actively commit to showing integrity). I'd say I'm long on personal/individual integrity and accountability (esp. if the same data can help me be a better writer).
Great breakdown and explanation Anna! It's really helpful to have this to learn more about these tools!
I do keep thinking about the watching the writing process happen. It's one of those things where with some folks, it's incredibly helpful to see and support. And also, in the wrong frame of mind, becomes harmful. And who gets to decide that and how is the thing that makes it challenging because it still often feels institutions and faculty by proxy get the power in that decision. It can feel too close to the "well, if you have nothing to hide, you should be fine with letting us search your car."
I have to wonder if Grammarly's next step is to create/step into the "plagiarism" market and be a competitor to TurnItIn...speculation--sure, but that projection certainly makes sense. If they make this work well for users (and normalize it as a practice), then sell the backend access to institutions, it would make $$$.
Thanks for the comment--I'm really happy the post seems helpful. Yeah, I appreciate your points and the comparison to police searches.
After some similar concerned comments on X, I am thinking some more about offering an alternative to process tracking: if they prefer not to share document history they could meet with me (or perhaps record a video) to chat about their writing process. Of course they might feel shy about doing that, and it's extra work, but I'm not sure what else could be an alternative...
I think Grammarly already offers plagiarism and AI detection. The profit incentives are definitely key to watch. I don't know much about Grammarly for Edu. I haven't worked anywhere that would consider paying for it, I don't think...
yeah--I think these tools definitely open up the exploration of process and evaluating processes to understand/appreciate outcome.
It makes me wonder what there might be to learn from other disciplines about how we do this (Math and "show your work" or Art--which of course, they also face that problem of capturing the process well)...
Mmm, great point. In writing instruction we do often assign reflections on the writing process, and that in itself is one of the strategies that helps some with deterring misuse. I'm still looking for the best ways to link this kind of process tracking with process reflection. It seems like it should be a native part of the process tracking report app that it nudges the author to reflect and annotate.
This is really interesting, Anna. Thanks for the breakdown.
I'm looking forward to playing with it. My hope is that it gives another option, beyond those AI detectors.
I'll be interested to see what other companies produce as alternatives, too.