1
00:00:00,200 --> 00:00:02,485
One message reaches your inbox.

2
00:00:02,485 --> 00:00:04,390
Another lands in spam.

3
00:00:04,390 --> 00:00:06,240
How does the computer decide?

4
00:00:06,625 --> 00:00:10,305
Easy. Someone told it which words to look
for.

5
00:00:10,708 --> 00:00:12,668
Sometimes. Rules can help.

6
00:00:12,668 --> 00:00:17,428
But a system can also learn patterns from
examples, and still make mistakes.

7
00:00:17,905 --> 00:00:22,973
I'm Layla. Today we'll make sense of AI:
what it does, how we got here,

8
00:00:22,973 --> 00:00:23,833
and what to check.

9
00:00:24,666 --> 00:00:26,656
Artificial intelligence is a broad field.

10
00:00:26,656 --> 00:00:30,082
Some systems use rules and knowledge
written by people.

11
00:00:30,082 --> 00:00:32,608
Others learn patterns from data.

12
00:00:32,608 --> 00:00:33,888
Some combine both.

13
00:00:34,361 --> 00:00:39,012
In this simplified example, a filter
learns from messages labeled spam or not

14
00:00:39,012 --> 00:00:43,892
spam. Later, it uses those learned
patterns to sort a new message.

15
00:00:44,292 --> 00:00:48,772
So it can put an invitation in the wrong
place, even after all that learning?

16
00:00:49,747 --> 00:00:53,439
Exactly. We need to know the task, the
evidence, and what happens when it

17
00:00:53,439 --> 00:00:54,287
gets something wrong.

18
00:00:54,801 --> 00:00:56,667
Keep those three questions close.

19
00:00:56,667 --> 00:01:01,874
They will help us look past the impressive
surface and see what a system can do.

20
00:01:01,874 --> 00:01:06,027
First, let's open the filter and work
through one decision together.

21
00:01:06,553 --> 00:01:09,272
Let's slow down that first decision.

22
00:01:09,272 --> 00:01:14,467
Imagine we write one rule: if the subject
contains the word free, send the

23
00:01:14,467 --> 00:01:15,697
message to spam.

24
00:01:16,083 --> 00:01:18,803
It catches a suspicious free prize.

25
00:01:18,883 --> 00:01:23,683
It also catches a perfectly good
invitation to a free library event.

26
00:01:23,783 --> 00:01:27,283
Our rule is doing exactly what we told it
to do.

27
00:01:27,667 --> 00:01:30,005
The library gets punished for being

28
00:01:30,005 --> 00:01:32,707
generous. I object.

29
00:01:33,250 --> 00:01:36,210
Objection accepted. We could add
exceptions.

30
00:01:36,210 --> 00:01:39,650
But every exception is another choice
someone has to maintain.

31
00:01:39,690 --> 00:01:43,810
For this example, let's try learning from
labeled messages instead.

32
00:01:44,416 --> 00:01:48,658
This approach is called machine learning:
learning patterns from data rather than

33
00:01:48,658 --> 00:01:51,104
writing every decision as a rule.

34
00:01:51,104 --> 00:01:54,988
Here, learning means adjusting a model
using examples.

35
00:01:54,988 --> 00:01:58,448
Our labels show which messages we wanted
in each tray.

36
00:01:58,833 --> 00:02:02,673
Now we hold back a message the model
hasn't used for learning, and try it.

37
00:02:02,703 --> 00:02:06,104
Getting old examples right is useful.

38
00:02:06,104 --> 00:02:08,833
Handling a new one is the test we care
about.

39
00:02:09,208 --> 00:02:13,528
So the lesson is bigger than memorizing
the answer sheet.

40
00:02:13,917 --> 00:02:16,610
Yes. And one good result isn't enough.

41
00:02:16,610 --> 00:02:20,077
We'd try many different messages and
inspect the mistakes.

42
00:02:20,173 --> 00:02:24,857
This tiny drawing explains the idea; it
isn't a measurement of

43
00:02:24,857 --> 00:02:25,997
a commercial filter.

44
00:02:26,375 --> 00:02:30,455
Who decides which messages count as spam
in those examples?

45
00:02:30,972 --> 00:02:32,893
In our demonstration, we do.

46
00:02:32,893 --> 00:02:35,784
And we need to be clear about what we
mean.

47
00:02:35,784 --> 00:02:38,993
Consider the same newsletter reaching two
people.

48
00:02:38,993 --> 00:02:40,353
One signed up for it.

49
00:02:40,353 --> 00:02:44,593
The other didn't. The words alone won't
explain that difference.

50
00:02:45,000 --> 00:02:49,777
Before judging our filter, we define the
job: which messages should this particular

51
00:02:49,777 --> 00:02:55,080
person receive? Then we choose examples
and labels that fit that job.

52
00:02:55,360 --> 00:02:58,600
A pile of examples doesn't choose our
purpose for us.

53
00:02:59,000 --> 00:03:02,164
So the human choices are part of the
story, even when the

54
00:03:02,164 --> 00:03:03,560
prediction is automatic.

55
00:03:04,092 --> 00:03:08,345
Your turn. A subject says, Free ticket
enclosed.

56
00:03:08,345 --> 00:03:11,670
Under our original rule, which tray gets
it?

57
00:03:11,670 --> 00:03:13,958
Take a moment before I answer.

58
00:03:16,958 --> 00:03:19,918
Spam, because the rule only checks that
word.

59
00:03:19,978 --> 00:03:23,625
Now a harder question: is the message
unwanted?

60
00:03:23,625 --> 00:03:26,318
We can't settle that from the subject
alone.

61
00:03:26,708 --> 00:03:29,828
I'd check who sent it, and whether I
expected a ticket.

62
00:03:30,208 --> 00:03:34,704
Good. You separated the system's output
from the truth of the situation.

63
00:03:34,704 --> 00:03:37,061
That's a small move with a long reach.

64
00:03:37,061 --> 00:03:40,768
We will use it again when an answer
arrives as a beautiful paragraph.

65
00:03:41,303 --> 00:03:45,700
You will hear the phrase narrow AI for
systems focused on one task or a small

66
00:03:45,700 --> 00:03:50,207
group of tasks. Other tools work across
many kinds of tasks.

67
00:03:50,583 --> 00:03:54,956
Researchers still debate what should count
as artificial general intelligence, or

68
00:03:54,956 --> 00:04:00,103
AGI. You don't need to settle that debate
to ask useful questions about a tool.

69
00:04:00,500 --> 00:04:02,140
Can it do the job you need?

70
00:04:02,140 --> 00:04:03,600
Under what conditions?

71
00:04:03,600 --> 00:04:06,740
How often does it fail in examples like
yours?

72
00:04:06,900 --> 00:04:11,540
A spectacular demonstration answers fewer
questions than it first appears to.

73
00:04:11,917 --> 00:04:16,237
If it says it loves my hat, should I take
the compliment?

74
00:04:16,829 --> 00:04:18,685
You may enjoy the compliment.

75
00:04:18,685 --> 00:04:23,265
But a sentence about a feeling isn't
evidence that the system experiences that

76
00:04:23,265 --> 00:04:29,185
feeling. What an AI can do is different
from whether it has feelings or awareness.

77
00:04:29,583 --> 00:04:33,103
In our classroom, characters have stories
and personalities.

78
00:04:33,156 --> 00:04:36,847
In the lesson, we'll focus on what we can
test and observe.

79
00:04:36,847 --> 00:04:40,223
Your hat remains beyond the scope of this
investigation.

80
00:04:40,832 --> 00:04:43,772
These ideas did not begin with a chat
window.

81
00:04:43,772 --> 00:04:48,047
In nineteen fifty-five, a research
proposal used the name artificial

82
00:04:48,047 --> 00:04:53,025
intelligence and proposed a summer meeting
at Dartmouth in nineteen fifty-six.

83
00:04:53,583 --> 00:04:58,063
Notice the two dates: the proposal, then
the planned project.

84
00:04:58,183 --> 00:05:02,863
A timeline should help us understand what
happened, not turn different events into

85
00:05:02,863 --> 00:05:04,463
one tidy birthday.

86
00:05:04,875 --> 00:05:08,235
I like seeing the work before the famous
moment.

87
00:05:08,625 --> 00:05:12,625
Me too. A proposal is a beginning, not a
finished machine.

88
00:05:12,645 --> 00:05:17,425
Let’s follow a few concrete milestones,
and ask what each one demonstrated.

89
00:05:17,902 --> 00:05:22,713
In nineteen ninety-seven, IBM's Deep Blue
defeated reigning world chess champion

90
00:05:22,713 --> 00:05:25,726
Garry Kasparov in a six-game match.

91
00:05:25,726 --> 00:05:29,193
It evaluated possible chess positions to
choose moves.

92
00:05:29,917 --> 00:05:31,597
That was an impressive achievement.

93
00:05:31,624 --> 00:05:36,677
It was also a specific test, with a board,
legal moves, and a clear result.

94
00:05:36,677 --> 00:05:39,117
Winning it was evidence about chess
performance.

95
00:05:39,500 --> 00:05:43,180
So the headline doesn't tell me whether it
can organize my desk.

96
00:05:43,583 --> 00:05:48,415
Exactly. Whenever a system succeeds, ask
which task the evidence covers.

97
00:05:48,415 --> 00:05:49,727
Keep the achievement.

98
00:05:49,727 --> 00:05:51,343
Keep its boundaries too.

99
00:05:51,889 --> 00:05:56,372
In twenty seventeen, the paper Attention
Is All You Need introduced the Transformer

100
00:05:56,372 --> 00:06:00,790
architecture. Attention helps connect
relevant parts of an input.

101
00:06:01,167 --> 00:06:04,820
Look at this little example: The ticket is
blue.

102
00:06:04,820 --> 00:06:10,627
It is on the desk. Understanding what it
refers to involves a relationship across

103
00:06:10,627 --> 00:06:15,207
the words. Our picture illustrates that
idea; it isn't a recording of

104
00:06:15,207 --> 00:06:16,447
a model's thoughts.

105
00:06:16,833 --> 00:06:20,113
The Transformer became an important design
for language models.

106
00:06:20,353 --> 00:06:23,353
AI can also use other designs.

107
00:06:23,353 --> 00:06:27,393
One diagram cannot explain every system
with AI on its label.

108
00:06:28,128 --> 00:06:31,152
ChatGPT launched in November twenty
twenty-two.

109
00:06:31,312 --> 00:06:35,952
A conversational interface gave people a
way to explore language-model capabilities

110
00:06:35,952 --> 00:06:37,312
by asking questions.

111
00:06:37,708 --> 00:06:40,428
Our timeline is a selection, not the whole
history.

112
00:06:40,428 --> 00:06:43,536
The useful pattern is that different
designs demonstrated different

113
00:06:43,536 --> 00:06:47,674
capabilities. So when a new tool appears,
we look at what changed, and how

114
00:06:47,674 --> 00:06:48,588
anyone checked it.

115
00:06:49,068 --> 00:06:53,460
Now put a confident paragraph beside our
misrouted invitation.

116
00:06:53,460 --> 00:06:57,712
Generative AI can produce false or
unsupported information that sounds

117
00:06:57,712 --> 00:07:01,160
plausible. This is often called
hallucination.

118
00:07:01,542 --> 00:07:04,982
Then how am I supposed to build anything
on top of it?

119
00:07:05,375 --> 00:07:07,475
By choosing the job and the checks
together.

120
00:07:07,475 --> 00:07:10,995
Imagine asking for three cheerful
invitation drafts.

121
00:07:10,995 --> 00:07:12,975
You can judge the tone yourself.

122
00:07:12,986 --> 00:07:18,148
Asking for the library's actual opening
time creates a different obligation:

123
00:07:18,148 --> 00:07:19,055
check the time.

124
00:07:19,458 --> 00:07:22,231
The same fluent voice can appear in both
answers.

125
00:07:22,231 --> 00:07:25,218
Style doesn't tell you which facts were
verified.

126
00:07:25,761 --> 00:07:30,075
Let's do the check. Our invented draft
says the event starts at one.

127
00:07:30,075 --> 00:07:33,945
The matching entry in our invented library
schedule says two.

128
00:07:34,025 --> 00:07:38,665
We check the event name and date as well,
so we don't compare two different events.

129
00:07:39,042 --> 00:07:43,442
Then I correct the invitation to two, and
keep the schedule with it.

130
00:07:44,333 --> 00:07:46,996
Yes. We have evidence for the correction.

131
00:07:46,996 --> 00:07:51,053
Asking the same chatbot, Are you sure,
would give us another response.

132
00:07:51,080 --> 00:07:53,373
It wouldn't replace opening the source.

133
00:07:53,750 --> 00:07:57,930
For something consequential, choose checks
strong enough for the consequence.

134
00:07:57,930 --> 00:08:02,390
A source must exist, be relevant, and
support the particular claim.

135
00:08:02,510 --> 00:08:06,150
A link that merely looks official is only
a starting point.

136
00:08:06,542 --> 00:08:10,942
And if I tell it the correct time, will it
remember tomorrow?

137
00:08:11,467 --> 00:08:13,693
That depends on the app and its settings.

138
00:08:13,693 --> 00:08:16,901
A new chat may not include your earlier
conversation.

139
00:08:16,901 --> 00:08:19,973
Some apps can save details or use past
chats.

140
00:08:20,375 --> 00:08:24,645
Supplying saved context is different from
teaching the underlying model a new fact

141
00:08:24,645 --> 00:08:29,855
through training. Check what the app says
it saves, what you can review, and what

142
00:08:29,855 --> 00:08:30,935
you can remove.

143
00:08:31,333 --> 00:08:35,493
So I should bring the schedule again if
the next task needs it.

144
00:08:35,875 --> 00:08:40,733
A good habit. Give the task the evidence
it needs, instead of assuming yesterday's

145
00:08:40,733 --> 00:08:42,835
correction followed you into the room.

146
00:08:43,389 --> 00:08:46,950
It also helps to distinguish a model from
the app around it.

147
00:08:46,950 --> 00:08:48,770
Think of our classroom project.

148
00:08:48,770 --> 00:08:52,983
We could use a model to draft wording,
while the app supplies the schedule and

149
00:08:52,983 --> 00:08:54,690
keeps the draft in a document.

150
00:08:55,083 --> 00:08:58,043
If that schedule is missing, the task has
changed.

151
00:08:58,150 --> 00:09:02,683
If the app adds a search tool, the
available actions have changed.

152
00:09:02,740 --> 00:09:07,563
Comparing two answers without noticing
those differences can be misleading.

153
00:09:07,958 --> 00:09:13,718
So when I ask what an AI can do, I should
say which setup I'm using.

154
00:09:14,125 --> 00:09:18,091
Yes. The model, the information it
receives, and the tools around it all

155
00:09:18,091 --> 00:09:19,405
belong in the description.

156
00:09:19,445 --> 00:09:21,965
Now we can make sense of the word agent.

157
00:09:22,582 --> 00:09:23,868
There's another distinction.

158
00:09:23,868 --> 00:09:26,748
An answer can describe what to do.

159
00:09:26,748 --> 00:09:32,095
An agentic system can use tools, inspect
results, and decide what step to take next

160
00:09:32,095 --> 00:09:33,815
while working toward a goal.

161
00:09:34,208 --> 00:09:37,888
In our example, the goal is to prepare an
invitation.

162
00:09:38,008 --> 00:09:40,648
A permitted tool reads the schedule.

163
00:09:40,648 --> 00:09:45,628
The system drafts the message, compares
the details, and stops for your review.

164
00:09:45,628 --> 00:09:48,208
Sending isn't part of this task.

165
00:09:48,583 --> 00:09:52,663
I want that boundary written down before
it touches my inbox.

166
00:09:53,042 --> 00:09:58,050
Yes. The tools, permissions, stopping
point, and checks belong in the design.

167
00:09:58,050 --> 00:10:02,434
More steps can create more opportunities
to help, and more places to

168
00:10:02,434 --> 00:10:03,362
make a mistake.

169
00:10:03,960 --> 00:10:05,430
Let's try one decision.

170
00:10:05,430 --> 00:10:09,238
The draft is ready, but its time disagrees
with the schedule.

171
00:10:09,238 --> 00:10:14,760
What should the system do next: send it,
inspect the schedule, or make the

172
00:10:14,760 --> 00:10:16,230
wording more impressive?

173
00:10:19,250 --> 00:10:20,754
Inspect the schedule.

174
00:10:20,754 --> 00:10:24,450
Beautiful words won't repair the wrong
time.

175
00:10:24,833 --> 00:10:28,833
Right. The next useful action is the one
that resolves the uncertainty.

176
00:10:28,873 --> 00:10:34,113
After the repair, we still review the
message and its recipient before sending.

177
00:10:34,193 --> 00:10:38,513
Finishing a draft and approving an action
are different checkpoints.

178
00:10:38,917 --> 00:10:41,477
What makes you excited about where this
could go?

179
00:10:41,943 --> 00:10:45,365
The possibility of more people making the
things they imagine.

180
00:10:45,365 --> 00:10:49,443
Someone getting help with a first draft,
testing an idea, or finding another way

181
00:10:49,443 --> 00:10:51,107
into a difficult lesson.

182
00:10:51,107 --> 00:10:52,755
That's the future I'm rooting for.

183
00:10:53,167 --> 00:10:54,447
It isn't a guarantee.

184
00:10:54,447 --> 00:10:59,567
What people build, who can use it, and how
carefully it is checked all matter.

185
00:10:59,627 --> 00:11:03,927
You can be curious about the possibilities
and demanding about the evidence at

186
00:11:03,927 --> 00:11:04,847
the same time.

187
00:11:05,449 --> 00:11:08,797
Before we leave, close the notes for three
questions.

188
00:11:08,797 --> 00:11:12,610
First: does every AI system learn from
data?

189
00:11:15,625 --> 00:11:20,393
No. Some use rules and knowledge written
by people, some learn patterns, and

190
00:11:20,393 --> 00:11:21,785
some combine approaches.

191
00:11:22,167 --> 00:11:25,447
Second: does a confident answer prove it
is correct?

192
00:11:28,458 --> 00:11:31,498
No. Check the claim against relevant
evidence.

193
00:11:31,498 --> 00:11:34,698
Our invitation sounded fine while its time
was wrong.

194
00:11:35,083 --> 00:11:39,243
Third: what extra question matters when an
AI can use tools to act?

195
00:11:42,250 --> 00:11:45,810
Ask what it is allowed to do, where it
must stop, and how you'll

196
00:11:45,810 --> 00:11:47,050
check the result.

197
00:11:47,290 --> 00:11:51,530
If you remembered part of that, you have
something solid to build on.

198
00:11:52,061 --> 00:11:56,737
Next time, we'll look more closely at how
computers learn: examples, adjustments,

199
00:11:56,737 --> 00:11:58,397
and the test of something new.

200
00:11:58,437 --> 00:12:03,507
For now, choose one AI tool and name its
task, its evidence, and one

201
00:12:03,507 --> 00:12:04,717
mistake you'd check.

202
00:12:05,125 --> 00:12:07,138
Rain doesn't fall the same way twice.

203
00:12:07,138 --> 00:12:08,805
Bring your questions.

204
00:12:08,845 --> 00:12:10,245
I'll see you next time.