Latest / Early Edition with Ryan Bridge / Kathryn McCallum: Canterbury University Associate Professor of Digital Education on teachers warn AI could replace human judgement in classrooms

Transcript

Kathryn McCallum: Canterbury University Associate Professor of Digital Education on teachers warn AI could replace human judgement in classrooms · Early Edition with Ryan Bridge

• 0:00
  Speaker 1: So there's a worry about AI in the classroom. The
• 0:02
  Speaker 1: PPTA is now saying to their teachers, keep control of
• 0:06
  Speaker 1: marking your students' work and don't be lured by artificial intelligence.
• 0:11
  Speaker 1: They say AI wouldn't actually lead to any improvement, especially
• 0:15
  Speaker 1: for teachers' development. Let's go to University of Canterbury Associate
• 0:19
  Speaker 1: Professor Catherine McCallum, who was with us this morning. Catherine,
• 0:22
  Speaker 1: good morning. Good morning. So what happens when a teacher
• 0:26
  Speaker 1: marks a kid's work and uses AI to do it?
• 0:30
  Speaker 1: Is there anything wrong?
• 0:33
  Speaker 2: No, no, there's nothing wrong with it. But it's about,
• 0:37
  Speaker 2: as you say, marking as a human activity. It's about
• 0:40
  Speaker 2: actually understanding the learner and the context. And so if
• 0:44
  Speaker 2: you're using AI, they don't understand that context. They don't
• 0:48
  Speaker 2: understand the learner. So it can look more efficient. It
• 0:52
  Speaker 2: can be speeding things up a bit. It's actually not,
• 0:55
  Speaker 2: you know, marking is a human activity that we need
• 0:59
  Speaker 2: to make sure that we're using AI effectively because it
• 1:02
  Speaker 2: doesn't pick up that nuance that, you know, marking should.
• 1:06
  Speaker 1: There's a great variance in what we're talking about here.
• 1:09
  Speaker 1: If you're a math teacher and there's a math equation,
• 1:11
  Speaker 1: there's only one answer. You know, surely that would be
• 1:14
  Speaker 1: self-evident or actually easy to mark. And then there's an
• 1:17
  Speaker 1: English essay, which might be five pages long.
• 1:21
  Speaker 2: Yes, yeah, definitely. So AI can be used in different ways.
• 1:26
  Speaker 2: And I think, as you say, for really simple assessments
• 1:29
  Speaker 2: where there isn't a lot of judgment, human judgment, I
• 1:33
  Speaker 2: think we could use AI to make that process quick.
• 1:37
  Speaker 2: But they're also using it so you don't actually understand
• 1:40
  Speaker 2: who your learner is. And so a lot of assessment
• 1:43
  Speaker 2: is about developing the learner. It's about giving them feedback.
• 1:46
  Speaker 2: And so... while you can use AI for feedback as
• 1:50
  Speaker 2: well it's about actually understanding where your learners are at
• 1:53
  Speaker 2: as well and so it's that nuance even in a
• 1:56
  Speaker 2: maths question you can use AI to speed it up
• 2:00
  Speaker 2: but it still needs to be that human process that's important.
• 2:04
  Speaker 1: What about the teachers union saying to teachers basically don't
• 2:07
  Speaker 1: do AI it's not good for you not good for
• 2:09
  Speaker 1: anyone is that a smart thing to say Well, I mean, I.
• 2:13
  Speaker 2: Think as everything with AI, there's nuance. So it's not
• 2:17
  Speaker 2: just it's a bad thing, but I think it's important
• 2:19
  Speaker 2: to understand what you're giving up when you're looking at
• 2:22
  Speaker 2: using AI. And it's not a blank black and white issue.
• 2:26
  Speaker 2: It's about what are you using it for, how you're
• 2:30
  Speaker 2: using it. And I think a lot of the problem
• 2:32
  Speaker 2: when you come to AI is seductive. And so the
• 2:38
  Speaker 2: issue around the teacher development is that it can actually
• 2:43
  Speaker 2: look like the answer's more correct because it comes across
• 2:47
  Speaker 2: as being more authoritative and it undermines teachers' confidence sometimes
• 2:51
  Speaker 2: if you aren't confident in that process. So I think
• 2:54
  Speaker 2: that's the problem. It's that AI looks good, it has information,
• 3:00
  Speaker 2: and it provides it in a very authoritative voice, but
• 3:03
  Speaker 2: what we need is a human judgment, and it can
• 3:05
  Speaker 2: undermine that.
• 3:07
  Speaker 1: All right.
• 3:07
  Speaker 2: Dr.
• 3:07
  Speaker 1: Catherine McKellen, thanks for your time this morning. University of Canterbury,
• 3:11
  Speaker 1: For more from Early Edition with Ryan Bridge, listen live
• 3:14
  Speaker 1: to News Talk ZB from 5 a.m. weekdays or follow
• 3:17
  Speaker 1: the podcast on iHeartRadio.