Which parts of your job would you LIKE to automate?

[marshwiggle]

The ChatGPT discussion made me curious about whether everyone has parts of their work that they would be glad to have done by a bot. If you think absolutely everything you do needs human judgement, I' d love to hear why.

For one thing, I'd be happy to have early alerts and flameout warnings automated, so that students who are AWOL or mostly so would get dealt with requiring no instructor checking, pleading, or listening to sob stories. Bot processing would have a consistent tone and process across the institution, and could automatically direct students to counseling, appeals processes, etc.

[sinenomine]

I'd be more than happy to let ChapGPT write my annual programmatic assessment reports.

[Caracal]

I'd be perfectly happy to have a bot grade reading response papers. I try to glance through a few of them before class because it can be useful as a starting point for discussion, but I always get behind on actually grading them. If there was a program that would just give check marks to the ones that are fine and flag ones that seem deficient for me to look at, that would be great.

[Sun_Worshiper]

Off the top of my head:

• Grading
• Creating essay prompts, templates, and scenarios for assignments
• Any and all bureaucratic tasks
• Assistance with research/writing, such as making tables, identifying the proper code for some particular statistical model, or even stretching sentences into paragraphs

I have used it for some of these things and it is quite helpful.

[mythbuster]

In addition to the call for grading help and academic alerts, I would love a bot to deal with emails requesting appointments etc. I fully realize that there are apps that allow students wo make appointments, but they often email me anyway even if you offer that up.

[Wahoo Redux]

The administration.

[Parasaurolophus]

Everything except collecting the paycheques.

More seriously, pretty much just marking. I enjoy the rest. I've already automated most of my marking, though.

[lightning]

peer review of journal articles (for the ones that I know will suck, after reading the abstract and skimming)

[Sun_Worshiper]

peer review of journal articles (for the ones that I know will suck, after reading the abstract and skimming)

That's a good one. Reviewing bad papers sucks (and some good ones) up a lot of my time these days.

[Parasaurolophus]

peer review of journal articles (for the ones that I know will suck, after reading the abstract and skimming)

That's a good one. Reviewing bad papers sucks (and some good ones) up a lot of my time these days.

I don't mind that. Those at institutions with majors or master's degrees or PhDs get the benefit of regularly working with and on cutting-edge research, developing new research programs, and, at the end of the process, reading a long article or monograph that does a decent job of summarizing the extant literature and breaking new ground. Purely by virtue of having students, let alone great ones. A big part of the reason my PhD supervisor knows so much and is so prolific is because of how many students he's had, and how many different corners of the subfield they explored with his help.

I get none of that, except via peer review. Even when the submission is godawful--and it usually isn't--it usually has something of value; a new idea/avenue of research, some references that are new to me... or even just the confidence boost of being able to identify what's going wrong, and the mental exercise of suggesting how to fix it.

I don't like getting bad papers, but I see them as an opportunity to exercise some professional muscles that otherwise don't get many outings. So, it feels like time well-spent to me!

[Sun_Worshiper]

peer review of journal articles (for the ones that I know will suck, after reading the abstract and skimming)

That's a good one. Reviewing bad papers sucks (and some good ones) up a lot of my time these days.

I don't mind that. Those at institutions with majors or master's degrees or PhDs get the benefit of regularly working with and on cutting-edge research, developing new research programs, and, at the end of the process, reading a long article or monograph that does a decent job of summarizing the extant literature and breaking new ground. Purely by virtue of having students, let alone great ones. A big part of the reason my PhD supervisor knows so much and is so prolific is because of how many students he's had, and how many different corners of the subfield they explored with his help.

I get none of that, except via peer review. Even when the submission is godawful--and it usually isn't--it usually has something of value; a new idea/avenue of research, some references that are new to me... or even just the confidence boost of being able to identify what's going wrong, and the mental exercise of suggesting how to fix it.

I don't like getting bad papers, but I see them as an opportunity to exercise some professional muscles that otherwise don't get many outings. So, it feels like time well-spent to me!

I agree that reviewing good papers helps one to stay current and part of the job is giving feedback to peers. But it would be tempting to let AI handle it if that was an option.

[ergative]

Everything to do with timetabling, attendance. We have an entire admin team, and somehow we're still the ones filling out spreadsheets about what our room size and AV needs are for a class that didn't run last year. Admin team, you can look it up from two years ago, can't you? Isn't that your job?

[Caracal]

Everything to do with timetabling, attendance. We have an entire admin team, and somehow we're still the ones filling out spreadsheets about what our room size and AV needs are for a class that didn't run last year. Admin team, you can look it up from two years ago, can't you? Isn't that your job?

In theory it would be nice if something could be my attendance for me. I always have to do it manually in class since I never have a TA to either take it or record the names on a sign up sheet. (I'm pretty slow at transferring data from one sheet to another and it's much more efficient for me to just take the time to do it in one place)

I suspect AI facial recognition would end up being a mess though even beyond the privacy concerns, and in the end wouldn't actually be more efficient and certainly wouldn't be cost efficient. I suspect this is true for lots of AI. There are some tasks it can probably do pretty well-I do suspect it could put checkmarks on my response papers-but in lots of cases it isn't actually going to make sense to replace humans with a bot. In the cases where it works it might just make humans more efficient. I'm not going to give a zero to a student on a response paper based on AI, but it could flag the ones I should take a look at and give checks to the rest. Even so, I would still probably need to spot check a few of the check marks just to make sure it's calibrated right.

[the_geneticist]

Everything to do with timetabling, attendance. We have an entire admin team, and somehow we're still the ones filling out spreadsheets about what our room size and AV needs are for a class that didn't run last year. Admin team, you can look it up from two years ago, can't you? Isn't that your job?

I second this!  And we're the ones who say "Dr. So-and-So isn't teaching [baskets 101] in fall, it's Dr. Other-Person, please update" since the admin team doesn't cross-check the teaching grid against the list of sabbaticals, new hires, etc.

[Sun_Worshiper]

Fyi, there are tools that automate attendance