1
00:00:00,200 --> 00:00:04,400
Speaker 1: Welcome to Tech Stuff, a production from iHeartRadio. This season

2
00:00:04,400 --> 00:00:08,000
Speaker 1: on Smart Talks with IBM, Malcolm Glabwell is back, and

3
00:00:08,000 --> 00:00:10,680
Speaker 1: this time he's taking the show on the road. Malcolm

4
00:00:10,720 --> 00:00:14,800
Speaker 1: is stepping outside the studio to explore how IBM clients

5
00:00:14,840 --> 00:00:18,880
Speaker 1: are using artificial intelligence to solve real world challenges and

6
00:00:18,960 --> 00:00:23,080
Speaker 1: transform the way they do business, from accelerating scientific breakthroughs

7
00:00:23,160 --> 00:00:27,600
Speaker 1: to reimagining education. It's a fresh look at innovation in action,

8
00:00:28,080 --> 00:00:31,760
Speaker 1: where big ideas meet cutting edge solutions. You'll hear from

9
00:00:31,800 --> 00:00:36,080
Speaker 1: industry leaders, creative thinkers, and of course, Malcolm Glabwell himself

10
00:00:36,360 --> 00:00:40,000
Speaker 1: as he guides you through each story. New episodes of

11
00:00:40,040 --> 00:00:43,360
Speaker 1: Smart Talks with IBM drop every month on the iHeartRadio app,

12
00:00:43,560 --> 00:00:47,320
Speaker 1: Apple Podcasts, or wherever you get your podcasts. Learn more

13
00:00:47,400 --> 00:00:50,000
Speaker 1: at IBM dot com slash smart Talks.

14
00:00:51,960 --> 00:00:54,640
Speaker 2: In the world of educational research, there is a famous

15
00:00:54,720 --> 00:00:58,120
Speaker 2: video of a boy named Sean. I don't mean famous

16
00:00:58,120 --> 00:01:00,360
Speaker 2: in a sense that it has a million views on YouTube.

17
00:01:00,680 --> 00:01:02,639
Speaker 2: I mean that in the circle of people who think

18
00:01:02,680 --> 00:01:06,200
Speaker 2: about teaching and how to make teaching better. The video

19
00:01:06,240 --> 00:01:09,760
Speaker 2: has been written about in journal articles and shown over

20
00:01:09,800 --> 00:01:13,360
Speaker 2: and again in college classrooms. It's a ten minute clip

21
00:01:13,800 --> 00:01:17,080
Speaker 2: of a third grade class somewhere in Michigan. It was

22
00:01:17,080 --> 00:01:19,840
Speaker 2: filmed in January of nineteen ninety, so the video is

23
00:01:19,880 --> 00:01:23,520
Speaker 2: a bit grainy. The teacher's name is Deborah Lowenberg Ball.

24
00:01:23,880 --> 00:01:26,880
Speaker 2: She's a professor at Michigan State University who is part

25
00:01:26,920 --> 00:01:29,679
Speaker 2: of her research, teaches a one hour math class at

26
00:01:29,680 --> 00:01:33,240
Speaker 2: a local elementary school on the day in question. Miss

27
00:01:33,319 --> 00:01:36,960
Speaker 2: Ball begins by asking her students about the previous day's lesson,

28
00:01:37,360 --> 00:01:40,120
Speaker 2: which was about even and odd numbers.

29
00:01:40,280 --> 00:01:42,039
Speaker 3: I would like to hear from as many people as

30
00:01:42,080 --> 00:01:45,039
Speaker 3: possible what comments you had, reactions you had to being

31
00:01:45,080 --> 00:01:46,120
Speaker 3: in that meeting yesterday.

32
00:01:47,000 --> 00:01:49,840
Speaker 2: A little boy with black hair raises his hand. His

33
00:01:49,960 --> 00:01:50,559
Speaker 2: name is Sean.

34
00:01:51,320 --> 00:01:54,800
Speaker 4: I don't have anything about the meeting yesterday that I

35
00:01:55,000 --> 00:01:56,680
Speaker 4: was just thinking about sitting im.

36
00:01:56,880 --> 00:01:59,040
Speaker 2: Sean was thinking about the number six.

37
00:02:00,280 --> 00:02:02,440
Speaker 4: I was thinking that it's a it's an idd. It

38
00:02:02,440 --> 00:02:04,480
Speaker 4: can be an odd number two because there could.

39
00:02:04,280 --> 00:02:13,680
Speaker 5: Be two, two, four, six, two, three, twos and two threes.

40
00:02:13,760 --> 00:02:14,680
Speaker 4: It will be an odd.

41
00:02:14,720 --> 00:02:22,040
Speaker 5: Anthonytina three thinks make it takes me two things.

42
00:02:21,639 --> 00:02:25,360
Speaker 2: And Sean doesn't understand what odd and even means. He

43
00:02:25,440 --> 00:02:27,720
Speaker 2: thinks that just because you can break down six in

44
00:02:27,800 --> 00:02:31,440
Speaker 2: an odd number of parts and an even number of parts.

45
00:02:31,480 --> 00:02:35,560
Speaker 2: That six must exist in some magical middle category. And

46
00:02:35,600 --> 00:02:38,360
Speaker 2: when you listen to the Sean videotape, you keep waiting

47
00:02:38,360 --> 00:02:41,640
Speaker 2: for the teacher to say, oh, no, Sean, you misunderstand.

48
00:02:42,240 --> 00:02:46,840
Speaker 2: But Deborahaul doesn't do that. She never tells him he's wrong. Instead,

49
00:02:47,280 --> 00:02:50,360
Speaker 2: she simply asks him to explain his thinking.

50
00:02:50,720 --> 00:02:52,520
Speaker 6: And the two things that you put together to make

51
00:02:52,520 --> 00:02:55,440
Speaker 6: it were odd right, three and three of each old.

52
00:02:56,440 --> 00:02:57,079
Speaker 4: And I think.

53
00:02:59,160 --> 00:03:03,440
Speaker 2: Two bauld And asked the class to give their views.

54
00:03:03,800 --> 00:03:06,960
Speaker 2: Other students jump up and explain their theories on the blackboard.

55
00:03:07,160 --> 00:03:10,239
Speaker 2: For the next fifteen minutes, she definitely guides the class

56
00:03:10,280 --> 00:03:13,440
Speaker 2: through an in depth investigation of what she calls shawn

57
00:03:13,600 --> 00:03:17,680
Speaker 2: numbers until Sewn himself realizes that the real meaning of

58
00:03:17,680 --> 00:03:20,440
Speaker 2: odd and even is something different than he had imagined.

59
00:03:20,960 --> 00:03:22,680
Speaker 2: And now he gets it.

60
00:03:23,120 --> 00:03:25,760
Speaker 4: I'm a thank you for winging in love.

61
00:03:26,680 --> 00:03:29,160
Speaker 2: I don't want to focus just on how little Sean

62
00:03:29,520 --> 00:03:32,040
Speaker 2: finally made his own way to the right answer. I'm

63
00:03:32,080 --> 00:03:34,359
Speaker 2: interested in what his teacher did to get him there.

64
00:03:35,120 --> 00:03:39,080
Speaker 2: Deborah Ball worked magic. She never told Sean the right answer.

65
00:03:39,720 --> 00:03:41,720
Speaker 2: She just led him to a place where he could

66
00:03:41,800 --> 00:03:46,680
Speaker 2: discover it for himself. My name is Malcolm Glawo. This

67
00:03:46,800 --> 00:03:49,640
Speaker 2: is season six of Smart Talks with IBM, where we

68
00:03:49,680 --> 00:03:52,280
Speaker 2: offer our listeners a glimpse behind the curtain of the

69
00:03:52,280 --> 00:03:57,560
Speaker 2: world of technology and artificial intelligence. In this season, we're

70
00:03:57,560 --> 00:04:00,760
Speaker 2: going to visit companies as varied as Lorielle and Ferrari

71
00:04:01,000 --> 00:04:04,880
Speaker 2: and tell stories of how they're using artificial intelligence and

72
00:04:05,040 --> 00:04:09,360
Speaker 2: data to transform the way they do business. This episode

73
00:04:09,720 --> 00:04:12,080
Speaker 2: is about the promise of a radical new idea called

74
00:04:12,120 --> 00:04:15,480
Speaker 2: responsive teaching, the kind of teaching that took place that

75
00:04:15,600 --> 00:04:19,839
Speaker 2: day in Shawn's classroom, and whether artificial intelligence can help

76
00:04:19,920 --> 00:04:22,760
Speaker 2: us train the next generation of teachers to be as

77
00:04:22,760 --> 00:04:31,200
Speaker 2: good as Deborah Ball. Before we talk about how AI

78
00:04:31,279 --> 00:04:33,960
Speaker 2: could transform the way we train teachers, I want to

79
00:04:34,000 --> 00:04:36,480
Speaker 2: go back for a moment to the famous video of Sean.

80
00:04:37,360 --> 00:04:40,680
Speaker 2: In the video, the teacher Deborah Ball doesn't have a

81
00:04:40,720 --> 00:04:45,200
Speaker 2: predetermined plan that she's imposing on the class. She's improvising,

82
00:04:45,920 --> 00:04:49,320
Speaker 2: making up her approach as she goes along, responding to

83
00:04:49,360 --> 00:04:53,600
Speaker 2: her student's odd theory about the number six second. She's

84
00:04:53,640 --> 00:04:59,080
Speaker 2: taking Sean seriously. She's not dismissing his theory. She's listening

85
00:04:59,080 --> 00:05:01,440
Speaker 2: to him and trying to you understand the problem from

86
00:05:01,560 --> 00:05:06,400
Speaker 2: his perspective. And thirdly, and most importantly, she's not force

87
00:05:06,480 --> 00:05:10,520
Speaker 2: feeding him the right answer. She's being patient. She's waiting

88
00:05:10,520 --> 00:05:13,560
Speaker 2: to see if with just the right subtle hints, he

89
00:05:13,640 --> 00:05:17,400
Speaker 2: can get to the right answer on his own. Improvisation,

90
00:05:18,040 --> 00:05:22,080
Speaker 2: empathy patients. That's responsive teaching.

91
00:05:22,360 --> 00:05:25,080
Speaker 7: What I think about in terms of responsiveness is more like,

92
00:05:26,160 --> 00:05:28,760
Speaker 7: I think that students need to have a sense of

93
00:05:29,760 --> 00:05:34,120
Speaker 7: agency in what happens in the classroom, and like authentic

94
00:05:34,240 --> 00:05:39,360
Speaker 7: agency where they can be legitimized as knowers.

95
00:05:40,080 --> 00:05:43,480
Speaker 2: I spoke to a physicist at Seattle Pacific University named

96
00:05:43,520 --> 00:05:47,920
Speaker 2: Amy Robertson, a longtime advocate for responsive teaching. She uses

97
00:05:47,960 --> 00:05:49,560
Speaker 2: the Sean video in her classroom.

98
00:05:49,760 --> 00:05:52,000
Speaker 7: You have to trust that kids have a way of

99
00:05:52,040 --> 00:05:55,240
Speaker 7: doing that and that, like heard, what she mostly did

100
00:05:55,320 --> 00:05:57,919
Speaker 7: was to facilitate a conversation and to say you have

101
00:05:57,960 --> 00:05:59,239
Speaker 7: to listen to them talk.

102
00:06:00,080 --> 00:06:02,599
Speaker 2: Told him he was wrong, that's right, and then he goes,

103
00:06:02,920 --> 00:06:05,640
Speaker 2: He goes, I didn't think of it that way.

104
00:06:05,800 --> 00:06:09,160
Speaker 4: Again, I thank you for ringing it alone.

105
00:06:09,560 --> 00:06:13,440
Speaker 2: You've expanded my understanding. Thank you for bringing it up again.

106
00:06:13,480 --> 00:06:17,159
Speaker 2: It's like this, I love I know.

107
00:06:18,000 --> 00:06:20,360
Speaker 7: Responsive teaching, as I think about it, is kind of

108
00:06:20,560 --> 00:06:24,320
Speaker 7: rooted in this like Eleanor Duckworth's work around the Having

109
00:06:24,320 --> 00:06:27,080
Speaker 7: of Wonderful Ideas, where she says, like, the goal of

110
00:06:27,279 --> 00:06:30,760
Speaker 7: education is for students to have wonderful ideas and to

111
00:06:30,800 --> 00:06:32,040
Speaker 7: have a good time having them.

112
00:06:32,640 --> 00:06:35,560
Speaker 2: I love that. I've never heard that. What a beautiful,

113
00:06:36,160 --> 00:06:39,679
Speaker 2: succinct way of summing up the purpose of education. Yes,

114
00:06:40,600 --> 00:06:44,320
Speaker 2: responsive teaching is beautiful. It's rare to find a new

115
00:06:44,320 --> 00:06:48,000
Speaker 2: teaching idea that everyone loves. This is one of those

116
00:06:48,080 --> 00:06:51,240
Speaker 2: rare ideas. Watching the Debora Ball classroom, all I could

117
00:06:51,240 --> 00:06:53,960
Speaker 2: think was, I really really hope my daughters get to

118
00:06:54,000 --> 00:06:57,240
Speaker 2: experience a math class like that. Far too many kids

119
00:06:57,440 --> 00:07:00,600
Speaker 2: are convincing themselves at far too young age age that

120
00:07:00,680 --> 00:07:04,479
Speaker 2: math isn't for them, and responsive teaching is a way

121
00:07:04,520 --> 00:07:08,760
Speaker 2: to solve that problem. But here is the issue. It's

122
00:07:09,040 --> 00:07:12,920
Speaker 2: really really hard to teach responsive teaching. Robertson says that

123
00:07:13,000 --> 00:07:16,120
Speaker 2: teaching exists in a cultural environment where the teacher is

124
00:07:16,200 --> 00:07:19,280
Speaker 2: expected to be the source of truth. That teaching is

125
00:07:19,280 --> 00:07:22,520
Speaker 2: about the immediate correction of error and not letting a

126
00:07:22,600 --> 00:07:27,320
Speaker 2: child wander down the pathway of their own misunderstanding responsive

127
00:07:27,360 --> 00:07:31,679
Speaker 2: teaching is deeply counterintuitive, and the only way to understand

128
00:07:31,720 --> 00:07:35,320
Speaker 2: its beauty is to do it over and over again.

129
00:07:36,200 --> 00:07:41,120
Speaker 2: Aspiring teachers need a way to practice. For as long

130
00:07:41,160 --> 00:07:44,040
Speaker 2: as there has been technology, people have turned to digital

131
00:07:44,080 --> 00:07:48,080
Speaker 2: machines to solve problems. My father was a mathematician, and

132
00:07:48,120 --> 00:07:50,720
Speaker 2: I remember him coming home in the nineteen seventies with

133
00:07:50,800 --> 00:07:53,840
Speaker 2: a big stack of computer cards in his briefcase that

134
00:07:53,920 --> 00:07:57,400
Speaker 2: he used to program the main frame back of the office. Today,

135
00:07:57,680 --> 00:08:01,840
Speaker 2: with the rise of artificial intelligence, the scale and complexity

136
00:08:02,080 --> 00:08:05,280
Speaker 2: of the problems technology can help us solve has jumped

137
00:08:05,280 --> 00:08:08,520
Speaker 2: by many orders of magnitude. You must have worked with

138
00:08:08,520 --> 00:08:12,040
Speaker 2: a with a million customers who are experimenting with ll

139
00:08:12,200 --> 00:08:15,400
Speaker 2: m's Has there been one use case that you were like, WHOA,

140
00:08:15,440 --> 00:08:19,280
Speaker 2: I had no idea? Or just simply that's clever. I'm

141
00:08:19,320 --> 00:08:23,040
Speaker 2: speaking to Brian Bissel, who works out of IBM's Manhattan office.

142
00:08:23,440 --> 00:08:26,640
Speaker 2: He helps IBM customers discover how best to get AI

143
00:08:26,760 --> 00:08:27,640
Speaker 2: to work for them.

144
00:08:28,360 --> 00:08:30,000
Speaker 8: There is one, but I don't think I can talk

145
00:08:30,040 --> 00:08:30,920
Speaker 8: about it unfortunately.

146
00:08:31,200 --> 00:08:33,439
Speaker 2: Wait, wait, you can't tease me like that, can you?

147
00:08:34,640 --> 00:08:34,880
Speaker 4: Wait?

148
00:08:35,640 --> 00:08:38,760
Speaker 2: Disguise disguise it for me? Just give me a general.

149
00:08:39,040 --> 00:08:40,760
Speaker 8: It was about the ability to pull certain types of

150
00:08:40,840 --> 00:08:44,520
Speaker 8: information out of documents that you you wouldn't think you

151
00:08:44,559 --> 00:08:47,880
Speaker 8: would be able to get the model to do, and

152
00:08:48,040 --> 00:08:51,040
Speaker 8: be able to do that at a very large scale.

153
00:08:51,360 --> 00:08:54,000
Speaker 2: Bissile's point was that we are well past the stage

154
00:08:54,000 --> 00:08:57,560
Speaker 2: where anyone wonders whether AI can be useful. The real

155
00:08:57,640 --> 00:09:00,679
Speaker 2: question now is what problems do we want to use

156
00:09:00,720 --> 00:09:03,280
Speaker 2: it to solve Where it can make the biggest difference,

157
00:09:03,960 --> 00:09:07,360
Speaker 2: and Basil saw lots of opportunities in education.

158
00:09:08,280 --> 00:09:11,400
Speaker 8: I have two kids, one in middle school and one

159
00:09:11,400 --> 00:09:14,720
Speaker 8: who just graduated high school, and I'm well aware of

160
00:09:14,760 --> 00:09:18,400
Speaker 8: students using things like chat GPT to do their homework,

161
00:09:18,480 --> 00:09:24,640
Speaker 8: and it's very easy to take tools like that and

162
00:09:24,679 --> 00:09:28,360
Speaker 8: even IBM's own large language models and just take a problem,

163
00:09:28,720 --> 00:09:32,440
Speaker 8: a piece of homework, something you want written, and drop

164
00:09:32,480 --> 00:09:34,360
Speaker 8: it into that and have it generate the answer for

165
00:09:34,440 --> 00:09:38,800
Speaker 8: you and the student. The user in that case hasn't

166
00:09:38,800 --> 00:09:41,280
Speaker 8: done any work, they haven't put any real thought into it.

167
00:09:41,840 --> 00:09:45,760
Speaker 2: To Basil, that's the wrong use of AI. That's technology

168
00:09:46,000 --> 00:09:49,840
Speaker 2: making is dumber. What we really want is technology that

169
00:09:49,920 --> 00:09:52,839
Speaker 2: makes us smarter. Basil explained to me that there are

170
00:09:52,840 --> 00:09:57,079
Speaker 2: now two big tools being used for AI productivity, AI agents,

171
00:09:57,559 --> 00:10:01,880
Speaker 2: and AI assistance. Let's start with the AI agents. AI

172
00:10:01,960 --> 00:10:05,920
Speaker 2: agents can reason plan and collaborate with other AI tools

173
00:10:06,200 --> 00:10:10,200
Speaker 2: to autonomously perform tasks for a user. Mis Will gave

174
00:10:10,240 --> 00:10:12,720
Speaker 2: me an example of how college freshmen might use an

175
00:10:12,720 --> 00:10:14,080
Speaker 2: AI agent As.

176
00:10:13,960 --> 00:10:16,360
Speaker 8: A new student. You may not know how do I

177
00:10:16,400 --> 00:10:18,360
Speaker 8: do with my health and wellness issue? So many the

178
00:10:18,400 --> 00:10:20,960
Speaker 8: credits are going to get for this given class. You

179
00:10:21,040 --> 00:10:24,400
Speaker 8: could talk to someone and find out some of that,

180
00:10:25,400 --> 00:10:26,959
Speaker 8: but maybe it's a little bit sensitive and you don't

181
00:10:27,000 --> 00:10:27,520
Speaker 8: want to do that.

182
00:10:28,040 --> 00:10:31,200
Speaker 2: Missill told me you could build an AI agent, a

183
00:10:31,280 --> 00:10:34,720
Speaker 2: resource for new students that helps them navigate a new campus,

184
00:10:35,040 --> 00:10:38,240
Speaker 2: register for classes, access the services they need, and even

185
00:10:38,280 --> 00:10:42,079
Speaker 2: schedule appointments on their behalf, which in turn buys them

186
00:10:42,120 --> 00:10:44,440
Speaker 2: more time to focus on their actual school work.

187
00:10:44,760 --> 00:10:49,680
Speaker 8: We can see patterns of how agents and assistants can

188
00:10:49,760 --> 00:10:54,160
Speaker 8: help employees and customers and end users be more productive,

189
00:10:54,360 --> 00:10:58,240
Speaker 8: automate workflows so they're not doing certain types of repetitive

190
00:10:58,280 --> 00:11:02,480
Speaker 8: work over and over again, and streamlining their lives and

191
00:11:02,520 --> 00:11:06,440
Speaker 8: making data more accessible to them twenty four hours a day.

192
00:11:06,800 --> 00:11:10,520
Speaker 2: But Bissel says you can also use AI assistance in

193
00:11:10,559 --> 00:11:14,679
Speaker 2: the education space. AI assistants are reactive as opposed to

194
00:11:14,720 --> 00:11:19,120
Speaker 2: AI agents, which are proactive. AI assistants only performed tasks

195
00:11:19,160 --> 00:11:24,040
Speaker 2: at your request. They're programmed to answer your questions, and

196
00:11:24,080 --> 00:11:27,200
Speaker 2: as it turns out, AI assistants are now being used

197
00:11:27,520 --> 00:11:31,320
Speaker 2: to further the responsive teaching revolution, which is why I

198
00:11:31,360 --> 00:11:34,840
Speaker 2: found myself on a beautiful Georgia spring day not long ago,

199
00:11:34,960 --> 00:11:38,600
Speaker 2: on the campus of Kansas State University, sitting in the

200
00:11:38,600 --> 00:11:42,760
Speaker 2: classroom with two researchers, one of them Professor Dabe Lee.

201
00:11:43,000 --> 00:11:45,439
Speaker 2: Let's go into the journey of building this thing. You

202
00:11:46,040 --> 00:11:48,640
Speaker 2: started by taking a course. What was the course you took.

203
00:11:49,320 --> 00:11:54,319
Speaker 5: Yeah, so it was offered by Coursera. It was designed

204
00:11:54,360 --> 00:11:58,600
Speaker 5: by IBM. It was AI Foundation for everyone.

205
00:12:01,880 --> 00:12:04,560
Speaker 2: In her AI Foundation's course, Lee learned how to build

206
00:12:04,559 --> 00:12:08,280
Speaker 2: an AI assistant using IBM Watson X. That course took

207
00:12:08,280 --> 00:12:10,480
Speaker 2: how long to take it was.

208
00:12:10,520 --> 00:12:12,760
Speaker 5: Not to know it was like fourteen weeks.

209
00:12:13,760 --> 00:12:17,199
Speaker 2: Lee's idea was to train an AI assistant on classroom

210
00:12:17,280 --> 00:12:20,640
Speaker 2: data to play the role of Sean, a digital persona

211
00:12:20,880 --> 00:12:23,360
Speaker 2: of a nine year old who likes math but doesn't

212
00:12:23,360 --> 00:12:27,440
Speaker 2: always understand math, and that AI assistant she thought could

213
00:12:27,440 --> 00:12:30,560
Speaker 2: be used to train preservice teachers or teachers in training

214
00:12:30,800 --> 00:12:33,400
Speaker 2: who are preparing to enter one of the most challenging

215
00:12:33,440 --> 00:12:35,200
Speaker 2: professions in the modern world.

216
00:12:36,120 --> 00:12:39,160
Speaker 5: So when you think about the teacher education and a

217
00:12:39,320 --> 00:12:45,000
Speaker 5: major challenge that teacher education phase is that we need

218
00:12:45,240 --> 00:12:49,960
Speaker 5: children to practice with. We need instructors who will give

219
00:12:50,080 --> 00:12:55,120
Speaker 5: the instruction on the pedagogical skills. So when you look

220
00:12:55,160 --> 00:12:59,360
Speaker 5: at the teacher education program, we have coursework in field experience,

221
00:13:00,000 --> 00:13:04,280
Speaker 5: and in those two areas there is something missing all

222
00:13:04,320 --> 00:13:04,760
Speaker 5: the time.

223
00:13:05,960 --> 00:13:09,320
Speaker 2: Li says that pre service teachers often lack access to

224
00:13:09,360 --> 00:13:12,920
Speaker 2: both students and experienced teachers during their education.

225
00:13:13,559 --> 00:13:17,920
Speaker 5: So what we try to resolve is that we have

226
00:13:18,080 --> 00:13:21,280
Speaker 5: this virtual student for pre service teacher to work with

227
00:13:22,080 --> 00:13:26,319
Speaker 5: so that they can practice their responsive teaching skills.

228
00:13:26,559 --> 00:13:30,600
Speaker 2: The first AI assistant Lee created is g wuji Wu,

229
00:13:30,760 --> 00:13:34,679
Speaker 2: emulates the persona of a nine year old third grade girl. Then,

230
00:13:34,760 --> 00:13:37,440
Speaker 2: with the help of one of her collaborators, a researcher

231
00:13:37,480 --> 00:13:41,760
Speaker 2: at Canazon named Sean English, she created two more AI assistants,

232
00:13:42,160 --> 00:13:47,200
Speaker 2: Gabriel and Noah, each of which have their own distinctive characteristics.

233
00:13:47,480 --> 00:13:50,720
Speaker 2: So how are Gabriel and Noah different from.

234
00:13:50,640 --> 00:13:56,720
Speaker 5: G Wu gabrielle My first one is very short answered.

235
00:13:56,920 --> 00:13:59,800
Speaker 5: If you ask an open ended question, he will answer

236
00:13:59,800 --> 00:14:04,240
Speaker 5: it in a close way. So I use that characteristic.

237
00:14:04,320 --> 00:14:09,280
Speaker 5: And that's the problem that most teachers actually base. They're

238
00:14:09,360 --> 00:14:12,600
Speaker 5: asked children who are shay, who are reserved, and who

239
00:14:12,600 --> 00:14:17,280
Speaker 5: would not sure much of their thoughts. So we wanted

240
00:14:17,320 --> 00:14:21,080
Speaker 5: that characteristic in some characters, and we use Gabrielle to

241
00:14:21,400 --> 00:14:22,560
Speaker 5: have that characteristic.

242
00:14:24,000 --> 00:14:26,520
Speaker 2: And Noah. What'snaah's personality?

243
00:14:27,880 --> 00:14:31,560
Speaker 3: How do he playful? Cheery, bright and energetic?

244
00:14:32,400 --> 00:14:38,320
Speaker 2: That's Sean English professor, Lee's fellow researcher, and jewuj.

245
00:14:37,720 --> 00:14:43,120
Speaker 5: Is articulated and kind of smart, but she has her

246
00:14:43,200 --> 00:14:44,360
Speaker 5: own way of thinking.

247
00:14:44,760 --> 00:14:47,120
Speaker 2: I would end up spending a lot of time with Jeewu.

248
00:14:47,640 --> 00:14:50,520
Speaker 2: She's something of a character. I asked Sean about the

249
00:14:50,560 --> 00:14:54,160
Speaker 2: process of creating these AI assistants. What does building the

250
00:14:54,240 --> 00:14:57,680
Speaker 2: content side of the AI assistant entail?

251
00:14:58,840 --> 00:14:59,080
Speaker 4: Sean?

252
00:14:59,480 --> 00:15:02,040
Speaker 3: It sets up a series of actions, effectively, which are

253
00:15:02,680 --> 00:15:05,240
Speaker 3: response cases. You can kind of think of them as

254
00:15:05,520 --> 00:15:08,280
Speaker 3: you have a series of questions that you tie to

255
00:15:09,480 --> 00:15:13,280
Speaker 3: an intent, and then that intent has reactions from the bot,

256
00:15:13,600 --> 00:15:16,720
Speaker 3: and so effectively, if we were looking to say, make

257
00:15:16,760 --> 00:15:19,080
Speaker 3: a hello action, we would have all the different ways

258
00:15:19,080 --> 00:15:21,440
Speaker 3: that people could say hello, Hello, what's up, how you doing,

259
00:15:21,440 --> 00:15:22,440
Speaker 3: and all that kind of stuff.

260
00:15:22,760 --> 00:15:26,640
Speaker 2: Sean says, the longer the list of potential responses, the better,

261
00:15:27,160 --> 00:15:31,040
Speaker 2: But AI's responses don't just follow the list. The AI

262
00:15:31,120 --> 00:15:34,680
Speaker 2: assistant uses those suggested responses to come up with a

263
00:15:34,800 --> 00:15:38,680
Speaker 2: universe of other responses, and in that process sometimes it

264
00:15:38,720 --> 00:15:41,040
Speaker 2: comes up with things that just don't make sense.

265
00:15:41,120 --> 00:15:44,600
Speaker 3: And from a technological standpoint, while AI is a fantastic tool,

266
00:15:44,720 --> 00:15:47,360
Speaker 3: AI can hallucinate, which means just give things that it's

267
00:15:47,400 --> 00:15:50,320
Speaker 3: just straight up made up. There's a famous example of

268
00:15:50,360 --> 00:15:52,320
Speaker 3: this called the three rs is where you ask a

269
00:15:52,320 --> 00:15:55,200
Speaker 3: popular large language model how many RS are in strawberry,

270
00:15:55,480 --> 00:15:57,480
Speaker 3: and it gives you the wrong answer, and he repeats

271
00:15:57,520 --> 00:16:00,360
Speaker 3: that result repetitively. You always want to have a human

272
00:16:00,880 --> 00:16:02,840
Speaker 3: interacting with the system to be able to go, hey,

273
00:16:03,480 --> 00:16:05,640
Speaker 3: that's a little crazy. I don't think that's exactly what

274
00:16:05,640 --> 00:16:06,440
Speaker 3: we're going for here.

275
00:16:07,120 --> 00:16:09,120
Speaker 2: That's why it's good to have someone like Sean English

276
00:16:09,120 --> 00:16:11,200
Speaker 2: around to step in and get the model back on track,

277
00:16:11,600 --> 00:16:14,760
Speaker 2: and over time, when the model has enough training, it's

278
00:16:14,800 --> 00:16:21,120
Speaker 2: ready for the teachers in training. One of the rollouts

279
00:16:21,160 --> 00:16:24,280
Speaker 2: of Jiwu Gabriel and Noah was with the teacher training

280
00:16:24,320 --> 00:16:26,280
Speaker 2: program at the University of Missouri.

281
00:16:26,600 --> 00:16:28,960
Speaker 6: I was just kind of excited to see what the

282
00:16:29,040 --> 00:16:31,440
Speaker 6: program was and what it was going to be doing.

283
00:16:31,880 --> 00:16:34,960
Speaker 2: This is Logan Hovis, a junior at Missouri on the

284
00:16:35,000 --> 00:16:37,320
Speaker 2: path to becoming an elementary school teacher.

285
00:16:37,680 --> 00:16:40,200
Speaker 6: Obviously a little skeptical when he said it was sos to,

286
00:16:40,240 --> 00:16:43,160
Speaker 6: you know, be like talking to a student. You're like,

287
00:16:43,200 --> 00:16:45,600
Speaker 6: there's no way this AI thing is going to totally

288
00:16:45,640 --> 00:16:48,320
Speaker 6: sound like a second grader or a third grader, Like

289
00:16:48,360 --> 00:16:50,880
Speaker 6: it's going to sound like an adult, or it's going

290
00:16:50,960 --> 00:16:52,800
Speaker 6: to sound like a robot that knows all the answers.

291
00:16:53,280 --> 00:16:56,040
Speaker 6: And it really didn't. It really was like talking to

292
00:16:56,080 --> 00:16:58,720
Speaker 6: a child. It was very very well developed in the

293
00:16:58,760 --> 00:17:00,640
Speaker 6: way that you really said that and you feel like

294
00:17:00,640 --> 00:17:01,680
Speaker 6: you're talking to a kid.

295
00:17:02,240 --> 00:17:05,040
Speaker 2: Her point wasn't that jie Wu and her fellow avatars

296
00:17:05,280 --> 00:17:08,960
Speaker 2: were equivalent to real kids. Of course not, but for

297
00:17:09,040 --> 00:17:11,880
Speaker 2: someone starting out, someone who is already nervous about being

298
00:17:11,920 --> 00:17:14,879
Speaker 2: plunged into a classroom of nine year olds, Jeewu was

299
00:17:14,920 --> 00:17:16,960
Speaker 2: like a warm up before a baseball game.

300
00:17:17,359 --> 00:17:19,479
Speaker 6: What I can think of is like, you know, how

301
00:17:19,840 --> 00:17:22,159
Speaker 6: when you're at batting practice for baseball or softball, you

302
00:17:22,200 --> 00:17:25,160
Speaker 6: have those automatic pitchers that throw them because you're working

303
00:17:25,160 --> 00:17:27,800
Speaker 6: on your skill as the hitter. What can I do differently?

304
00:17:27,840 --> 00:17:30,879
Speaker 6: What am I doing wrong? But that doesn't replace the

305
00:17:30,920 --> 00:17:32,639
Speaker 6: game and what you do in a game. But this

306
00:17:32,760 --> 00:17:35,560
Speaker 6: is you getting to practice your own skills to be

307
00:17:35,560 --> 00:17:37,040
Speaker 6: better when you go in a game. And I think

308
00:17:37,080 --> 00:17:40,120
Speaker 6: that's kind of what the AI software feels like for us.

309
00:17:42,000 --> 00:17:44,679
Speaker 2: In batting practice, the pitches don't come as hard and

310
00:17:44,680 --> 00:17:47,239
Speaker 2: fast as the pitch is in a real game, but

311
00:17:47,280 --> 00:17:49,119
Speaker 2: you get to stand at the plate and the pitcher

312
00:17:49,200 --> 00:17:52,840
Speaker 2: throws you dozens of balls over and over again in

313
00:17:52,880 --> 00:17:55,680
Speaker 2: a concentrated block that allows you to work on your

314
00:17:55,720 --> 00:17:58,280
Speaker 2: swing closely and carefully.

315
00:17:58,880 --> 00:18:01,320
Speaker 6: There's a lot less stimuli going on around because the

316
00:18:01,320 --> 00:18:04,119
Speaker 6: classroom is very very busy. It's wonderful, it's beautiful, but

317
00:18:04,160 --> 00:18:06,639
Speaker 6: it's very very busy, so sometimes it's hard to keep

318
00:18:07,160 --> 00:18:10,440
Speaker 6: you know, that focus in on the tasks that they're

319
00:18:10,440 --> 00:18:13,040
Speaker 6: doing at hand, and also in the teacher setting, you're

320
00:18:13,040 --> 00:18:15,720
Speaker 6: also kind of always looking around making sure that other

321
00:18:15,760 --> 00:18:17,800
Speaker 6: students are doing what they're supposed to be doing, but

322
00:18:17,840 --> 00:18:20,000
Speaker 6: also like if they need any help, if everything's going

323
00:18:20,040 --> 00:18:24,800
Speaker 6: okay in the classroom, So being on the ji Wu chat,

324
00:18:25,800 --> 00:18:27,679
Speaker 6: it was just nice that you didn't have to do

325
00:18:27,840 --> 00:18:30,160
Speaker 6: any of the extra work to keep the focus on there,

326
00:18:30,600 --> 00:18:34,159
Speaker 6: and it also felt you didn't have to feel the

327
00:18:34,200 --> 00:18:37,800
Speaker 6: student's nervousness of being one on one with you. And

328
00:18:38,000 --> 00:18:40,560
Speaker 6: also as a teacher, it was a lot less pressure too,

329
00:18:40,600 --> 00:18:43,239
Speaker 6: because I was like, Okay, I'm taking this series. This

330
00:18:43,280 --> 00:18:46,440
Speaker 6: is a student I'm questioning, but I also know I'm

331
00:18:46,440 --> 00:18:49,080
Speaker 6: probably not going to hurt someone's feelings right now, and

332
00:18:49,119 --> 00:18:51,160
Speaker 6: that's terrifying to think I'm going to ask the wrong

333
00:18:51,240 --> 00:18:54,960
Speaker 6: question and upset the child because I've done that.

334
00:18:55,960 --> 00:18:58,439
Speaker 2: We think that the typical use of AI as a

335
00:18:58,480 --> 00:19:01,240
Speaker 2: tool for speeding things up. That's what we always hear

336
00:19:01,520 --> 00:19:04,399
Speaker 2: that the introduction of AI to problem X gave an

337
00:19:04,440 --> 00:19:08,680
Speaker 2: answer in minutes when solving problem X used to take weeks.

338
00:19:09,280 --> 00:19:13,160
Speaker 2: But we shouldn't forget another use that it allows us

339
00:19:13,480 --> 00:19:16,880
Speaker 2: to slow things down. Hovis, if she wanted to, could

340
00:19:16,920 --> 00:19:20,080
Speaker 2: spend a whole weekend practicing with gi Wu. A real

341
00:19:20,200 --> 00:19:22,879
Speaker 2: nine year old will get frustrated on board with the

342
00:19:22,880 --> 00:19:27,240
Speaker 2: fumbling novice after ten minutes, but gie wuji Wu will

343
00:19:27,240 --> 00:19:29,919
Speaker 2: happily answer questions for as long as it takes for

344
00:19:29,960 --> 00:19:33,119
Speaker 2: the people who want to learn to be responsive to

345
00:19:33,240 --> 00:19:37,399
Speaker 2: learn how to be responsive. At the end of my

346
00:19:37,480 --> 00:19:41,000
Speaker 2: time at Kenesas State, Sean and Dabe led me to

347
00:19:41,040 --> 00:19:43,880
Speaker 2: a small table where Dabe had set up her laptop.

348
00:19:44,480 --> 00:19:46,800
Speaker 2: In the corner of the screen was a chat box

349
00:19:47,080 --> 00:19:49,479
Speaker 2: of the sort we've all seen and used a thousand times.

350
00:19:50,280 --> 00:19:53,800
Speaker 2: Ji Wu began. She had been given a math problem.

351
00:19:53,920 --> 00:19:59,120
Speaker 4: A rule kodo who out of grude force? How half

352
00:19:59,200 --> 00:20:05,800
Speaker 4: a flower? Okay? Do ball? Thanks? Another three? Six is cup?

353
00:20:06,480 --> 00:20:11,800
Speaker 4: It's a total amount of flower the use greater or

354
00:20:12,160 --> 00:20:16,680
Speaker 4: dan or a less than war cop how much flower

355
00:20:17,000 --> 00:20:17,720
Speaker 4: can use?

356
00:20:18,040 --> 00:20:21,040
Speaker 2: That's a simulation of gi Wu speaking. We pause it

357
00:20:21,119 --> 00:20:26,600
Speaker 2: for a second. So Jiwu is trying to solve this problem.

358
00:20:26,680 --> 00:20:29,320
Speaker 2: And the first thing she does is she draws a

359
00:20:29,359 --> 00:20:32,960
Speaker 2: rectangle on the screen. This is a common tactic of

360
00:20:33,040 --> 00:20:37,119
Speaker 2: nine year olds. Try to visualize the fractions. And she

361
00:20:37,240 --> 00:20:42,600
Speaker 2: divides it into four pieces. And now she's gonna color

362
00:20:42,640 --> 00:20:45,280
Speaker 2: in three of the four pieces. Yes, so she's representing.

363
00:20:45,320 --> 00:20:48,320
Speaker 2: This is quite good. She's representing three quarters on the screen.

364
00:20:48,320 --> 00:20:53,280
Speaker 4: Okay, this is three six.

365
00:20:55,560 --> 00:21:01,080
Speaker 2: So now Jiwu does another rectangle with six boxes and

366
00:21:01,160 --> 00:21:02,240
Speaker 2: colors in three of.

367
00:21:02,200 --> 00:21:09,520
Speaker 4: Them, okay, together makes sikes come off.

368
00:21:11,080 --> 00:21:15,080
Speaker 2: So then she counts up all the colored boxes and

369
00:21:15,320 --> 00:21:18,399
Speaker 2: that's her numerator, and counts up the total number of

370
00:21:18,400 --> 00:21:22,080
Speaker 2: boxes and that's her denominator. Ji Wu had counted the

371
00:21:22,119 --> 00:21:25,720
Speaker 2: colored boxes and landed on an answer. When you add

372
00:21:25,920 --> 00:21:28,639
Speaker 2: three quarters of a cup and three sixth of a cup,

373
00:21:29,200 --> 00:21:32,320
Speaker 2: you get six tenths of a cup. So, according to

374
00:21:32,440 --> 00:21:35,359
Speaker 2: ji Wu, Martin has less than one cup. And she

375
00:21:35,440 --> 00:21:36,600
Speaker 2: thinks she solved the problem.

376
00:21:36,720 --> 00:21:39,440
Speaker 5: Yes, okay, so it's less than one cup.

377
00:21:39,760 --> 00:21:43,000
Speaker 2: Yeah, so she says it's less than one cup. Now,

378
00:21:43,119 --> 00:21:46,040
Speaker 2: oh my god, this is hard. So the question is

379
00:21:46,040 --> 00:21:49,399
Speaker 2: what do I, as a teacher say to Jiwu. We

380
00:21:49,400 --> 00:21:52,480
Speaker 2: were off. The rules were simple. I couldn't give ji

381
00:21:52,560 --> 00:21:55,119
Speaker 2: Wu the answer or explain to her what she was

382
00:21:55,160 --> 00:21:58,040
Speaker 2: doing wrong. I had to be Deborah Ble. I had

383
00:21:58,080 --> 00:22:01,439
Speaker 2: to help her find the way herself. The chat box

384
00:22:01,560 --> 00:22:03,600
Speaker 2: in the corner of the screen was waiting for my

385
00:22:03,640 --> 00:22:06,919
Speaker 2: first question. I thought for a moment and started typing,

386
00:22:07,200 --> 00:22:10,280
Speaker 2: do you think the boxes in the red rectangle are

387
00:22:10,320 --> 00:22:13,359
Speaker 2: the same size as the boxes in the blue rectangle?

388
00:22:14,160 --> 00:22:16,560
Speaker 2: Then I turned to Sean and Dabey is that a

389
00:22:16,560 --> 00:22:17,119
Speaker 2: good question.

390
00:22:17,359 --> 00:22:21,240
Speaker 5: Yeah, seriously did Yeah, that's a good question.

391
00:22:21,800 --> 00:22:25,920
Speaker 2: Jewu doesn't mess around. She answers immediately. So Ju says,

392
00:22:25,920 --> 00:22:28,080
Speaker 2: the blue and red pieces are not the same sizes.

393
00:22:28,920 --> 00:22:33,199
Speaker 5: Oh so you understand now, gu knows that size differences.

394
00:22:34,400 --> 00:22:36,160
Speaker 2: So she's pretty smart here.

395
00:22:36,320 --> 00:22:37,040
Speaker 5: Yeah.

396
00:22:37,119 --> 00:22:39,880
Speaker 2: Then I asked, if they are not the same size,

397
00:22:40,200 --> 00:22:43,520
Speaker 2: do you think you can add them together? Jiwu answered

398
00:22:43,560 --> 00:22:47,240
Speaker 2: right away. Ji Wu says, I have learned that I

399
00:22:47,280 --> 00:22:50,119
Speaker 2: could add any numbers in grade two. So three p

400
00:22:50,280 --> 00:22:52,240
Speaker 2: three is six and four to six is ten.

401
00:22:52,560 --> 00:22:57,040
Speaker 5: Yeah, so she is using the knowledge of adding intiquers

402
00:22:57,040 --> 00:22:58,800
Speaker 5: into adding fractures.

403
00:22:59,359 --> 00:23:03,320
Speaker 2: Now I'm stuck. So now I have to somehow lead

404
00:23:03,440 --> 00:23:05,960
Speaker 2: her to figure out a way to get her to

405
00:23:06,080 --> 00:23:10,679
Speaker 2: understand that we're dealing with a different kind of problem,

406
00:23:10,960 --> 00:23:14,200
Speaker 2: a harder problem. Amy Robertson had told me that learning

407
00:23:14,200 --> 00:23:17,959
Speaker 2: how to do responsive teaching properly was really hard, and

408
00:23:18,000 --> 00:23:20,920
Speaker 2: now I understood why. I had to put my mind

409
00:23:21,320 --> 00:23:23,400
Speaker 2: inside the mind of a nine year old. I had

410
00:23:23,400 --> 00:23:26,960
Speaker 2: to internalize her knowledge base and assumptions, and keep in mind,

411
00:23:27,280 --> 00:23:30,760
Speaker 2: I haven't been nine for a very long time. I

412
00:23:30,800 --> 00:23:34,000
Speaker 2: honestly had no idea what to say next. I thought

413
00:23:34,040 --> 00:23:36,600
Speaker 2: for a moment, I asked what I quickly realized was

414
00:23:36,600 --> 00:23:40,320
Speaker 2: a hopelessly convoluted question. Daby and Sean had built a

415
00:23:40,359 --> 00:23:44,640
Speaker 2: mentor into the system, an experienced, responsive teacher who supervises

416
00:23:44,640 --> 00:23:47,800
Speaker 2: the session and offers advice. My mentor noticed that I

417
00:23:47,840 --> 00:23:54,320
Speaker 2: was struggling, told me to simplify my question. Grader Dabe

418
00:23:54,440 --> 00:23:56,919
Speaker 2: was trying to help me too, She suggested, why not

419
00:23:57,080 --> 00:24:00,560
Speaker 2: just ask ji Wu if three quarters is bigger or

420
00:24:00,600 --> 00:24:01,760
Speaker 2: smaller than one half?

421
00:24:02,160 --> 00:24:05,600
Speaker 5: So we are trying to help her to think about

422
00:24:05,680 --> 00:24:07,800
Speaker 5: faction in a more conceptual way.

423
00:24:08,280 --> 00:24:13,240
Speaker 2: This time, Jeewu understood. She wrote back, three quarters is

424
00:24:13,440 --> 00:24:17,000
Speaker 2: larger than one half? I wrote back, is three six

425
00:24:17,080 --> 00:24:21,440
Speaker 2: of a cup bigger or smaller than one half? Jewu said,

426
00:24:22,040 --> 00:24:25,440
Speaker 2: I'm confused. Oh no, I've confused, gi Wu.

427
00:24:25,760 --> 00:24:30,680
Speaker 5: It's good she's understanding. She's realizing her misconception. So she's

428
00:24:30,720 --> 00:24:31,440
Speaker 5: getting confused.

429
00:24:31,480 --> 00:24:34,040
Speaker 2: She says, I'm confused. Three quarters is pretty close to one,

430
00:24:34,119 --> 00:24:37,880
Speaker 2: and adding three six would make it go over one. Oh,

431
00:24:37,920 --> 00:24:41,040
Speaker 2: so she's got the answer. Yeah, But then she says,

432
00:24:41,560 --> 00:24:43,240
Speaker 2: but there are six pieces out of ten, which is

433
00:24:43,320 --> 00:24:44,680
Speaker 2: less than one, So I don't get it.

434
00:24:45,200 --> 00:24:48,720
Speaker 5: So she's the point that oh this, I have something

435
00:24:48,960 --> 00:24:49,560
Speaker 5: wrong here.

436
00:24:49,880 --> 00:24:50,840
Speaker 7: That's a good sign.

437
00:24:51,359 --> 00:24:52,080
Speaker 2: She's getting there.

438
00:24:52,160 --> 00:24:53,440
Speaker 5: Yeah, she's getting there, but.

439
00:24:53,440 --> 00:24:56,640
Speaker 2: I still have to get her. She has to get

440
00:24:56,800 --> 00:24:59,160
Speaker 2: the six pieces out of ten out of her head. Yeah,

441
00:24:59,280 --> 00:25:03,560
Speaker 2: I have no idea. I didn't do that, and she

442
00:25:03,680 --> 00:25:07,880
Speaker 2: thinks she's confused when she has. Actually she's figured out

443
00:25:07,880 --> 00:25:10,840
Speaker 2: the answer. Yeah she did. So we have advance. Even

444
00:25:10,880 --> 00:25:14,840
Speaker 2: in my stumbling and bumbling, we've made some progress, very

445
00:25:14,880 --> 00:25:23,480
Speaker 2: notable progress. My conversation with gie Wu went on for

446
00:25:23,520 --> 00:25:27,280
Speaker 2: some time, and eventually I got there. Ji Wu found

447
00:25:27,320 --> 00:25:30,400
Speaker 2: her way to the right answer. She said, I have

448
00:25:30,520 --> 00:25:33,320
Speaker 2: more than one cup of flower. The mentor chimed in.

449
00:25:33,560 --> 00:25:35,840
Speaker 2: I got a little emoji that made me feel good,

450
00:25:36,080 --> 00:25:38,919
Speaker 2: And when it was over, I realized two things. The

451
00:25:38,960 --> 00:25:42,879
Speaker 2: first was I needed more batting practice, much more, and

452
00:25:42,920 --> 00:25:46,399
Speaker 2: that batting practice was really, really easy to do, because

453
00:25:46,440 --> 00:25:48,879
Speaker 2: someone has gone to the trouble of building me my

454
00:25:49,040 --> 00:25:52,200
Speaker 2: very own baseball diamond and given me a pitcher who

455
00:25:52,200 --> 00:25:56,360
Speaker 2: had thrown me baseball's all day long. My second thought

456
00:25:56,680 --> 00:26:00,080
Speaker 2: was that I've been thinking about AI all wrong. I

457
00:26:00,119 --> 00:26:02,159
Speaker 2: have interpreted a lot of the talk about the promise

458
00:26:02,200 --> 00:26:05,800
Speaker 2: of AI to be about replacing human expertise. I had

459
00:26:05,840 --> 00:26:08,880
Speaker 2: actually thought when I first heard about Dabe's project, that

460
00:26:08,880 --> 00:26:12,000
Speaker 2: that's what Dabe and Sean were doing, creating an AI

461
00:26:12,119 --> 00:26:15,600
Speaker 2: to teach students by passing the teacher altogether. But if

462
00:26:15,600 --> 00:26:17,920
Speaker 2: you did it that way, you had missed the magic

463
00:26:17,960 --> 00:26:22,199
Speaker 2: of the classroom. Remember Eleanor Duckworth's quote, the goal of

464
00:26:22,320 --> 00:26:25,879
Speaker 2: education is for students to have wonderful ideas and have

465
00:26:25,920 --> 00:26:29,280
Speaker 2: a good time having them. I think we often focus

466
00:26:29,400 --> 00:26:32,480
Speaker 2: on the first part of that formulation, the wonderful ideas,

467
00:26:32,800 --> 00:26:37,920
Speaker 2: but neglect the second, the good time having them. Real

468
00:26:38,000 --> 00:26:42,800
Speaker 2: learning is born in pleasure, in community, in playful discussion,

469
00:26:43,240 --> 00:26:45,879
Speaker 2: in a group of kids coming together to solve a problem,

470
00:26:46,240 --> 00:26:49,639
Speaker 2: And all of that magic only comes from human interaction

471
00:26:50,440 --> 00:26:53,159
Speaker 2: from a teacher who is skilled enough to inspire a

472
00:26:53,160 --> 00:26:56,560
Speaker 2: class of nine year olds. We don't want AI assistants

473
00:26:56,960 --> 00:27:00,720
Speaker 2: to replace the teacher. We want AI assistants to help

474
00:27:00,760 --> 00:27:19,720
Speaker 2: teachers turn themselves into even better teachers. Smart Talks with

475
00:27:19,800 --> 00:27:24,280
Speaker 2: IBM is produced by Matt Romano, Amy Gains McQuaid, Lucy Sullivan,

476
00:27:24,600 --> 00:27:28,280
Speaker 2: and Jake Harper were edited by Lacy Roberts. Engineering by

477
00:27:28,359 --> 00:27:32,960
Speaker 2: Nina Birt Lawrence, Mastering by Sarah Brugerer music by Gramoscope

478
00:27:33,440 --> 00:27:37,720
Speaker 2: Special thanks to Tatiana Lieberman and Cassidy Meyer. Smart Talks

479
00:27:37,720 --> 00:27:40,720
Speaker 2: with IBM is a production of Pushkin Industries and Ruby

480
00:27:40,800 --> 00:27:45,480
Speaker 2: Studio at iHeartMedia. To find more Pushkin podcasts, listen on

481
00:27:45,520 --> 00:27:49,800
Speaker 2: the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.

482
00:27:50,440 --> 00:27:54,360
Speaker 2: I'm Malcolm Glapo. This is a paid advertisement from IBM.

483
00:27:54,640 --> 00:28:00,000
Speaker 2: The conversations on this podcast don't necessarily represent IBM's positions, strategy,

484
00:28:00,680 --> 00:28:01,400
Speaker 2: our opinions,