What is Generative AI—and why does it matter for education?
In this AI Foundations video from Ed3, we break down what generative AI actually is, how it differs from traditional machine learning, and why understanding it is essential for educators, school leaders, and anyone working with students today.
Most AI systems analyze or predict. Generative AI creates.It produces new text, images, audio, and video by remixing patterns learned from massive datasets. While the results may feel creative, generative AI is ultimately a powerful form of pattern prediction—not independent thinking.
This video explains:
A common misconception is that generative AI produces “original ideas.” It doesn’t. It simulates originality by recombining existing patterns. That distinction matters—especially when we’re teaching students about learning, creativity, authorship, and digital citizenship.
Ed3’s approach focuses on using AI as a thinking partner, not a replacement for student thinking, and on redesigning learning experiences so understanding, process, and reflection matter more than the final product.
This video is part of the AI Foundations series by Ed3, supporting educators worldwide in making informed, ethical, and human-centered decisions about AI in classrooms.
👉 Learn more about Ed3: https://www.ed3global.org
👉 Explore professional learning, courses, and events designed for educators navigating AI responsibly.
👉 Join our community of practice: https://community.ed3global.org
00:00 What makes generative AI different
00:22 How generative AI creates content
00:45 AI, machine learning, and deep learning explained
01:24 Why AI outputs feel creative
01:48 Copyright, authorship, and digital citizenship
02:17 Why banning AI doesn’t work
02:32 Redesigning assignments for ethical AI use
03:05 Four strategies to support students with generative AI
03:58 Separating AI hype from classroom reality
1
00:00:09,900 --> 00:00:11,166
Most AI systems
2
00:00:11,166 --> 00:00:13,633
analyze, categorize, or predict.
3
00:00:13,766 --> 00:00:16,466
Generative AI does something different.
4
00:00:16,466 --> 00:00:17,900
It creates.
5
00:00:22,000 --> 00:00:23,733
Generative AI refers to models
6
00:00:23,733 --> 00:00:24,966
that can produce new content
7
00:00:24,966 --> 00:00:26,366
like text, images,
8
00:00:26,366 --> 00:00:27,833
audio, or video
9
00:00:27,833 --> 00:00:30,433
that resemble human created work.
10
00:00:30,433 --> 00:00:32,066
Unlike traditional machine learning,
11
00:00:32,066 --> 00:00:34,166
which classifies or predicts
12
00:00:34,166 --> 00:00:35,966
generative AI, uses deep learning
13
00:00:35,966 --> 00:00:37,866
models trained on massive data sets
14
00:00:37,866 --> 00:00:40,033
to generate outputs that seem novel
15
00:00:40,033 --> 00:00:41,300
but are shaped by the patterns
16
00:00:41,300 --> 00:00:42,833
in its training data.
17
00:00:42,833 --> 00:00:45,400
Generative AI sits within deep learning,
18
00:00:45,400 --> 00:00:47,333
which sits within machine learning,
19
00:00:47,333 --> 00:00:48,433
which then sits
20
00:00:48,433 --> 00:00:50,900
within artificial intelligence.
21
00:00:50,900 --> 00:00:52,200
Generative AI uses
22
00:00:52,200 --> 00:00:53,666
billions of pieces of data
23
00:00:53,666 --> 00:00:55,233
to remix content.
24
00:00:55,233 --> 00:00:56,700
It combines fragments of what
25
00:00:56,700 --> 00:00:57,466
it's learned into
26
00:00:57,466 --> 00:00:58,533
something that sounds
27
00:00:58,533 --> 00:00:59,733
or looks new.
28
00:00:59,733 --> 00:01:00,733
To humans,
29
00:01:00,733 --> 00:01:02,433
that looks like creativity,
30
00:01:02,433 --> 00:01:04,466
but it's actually very advanced
31
00:01:04,466 --> 00:01:05,966
pattern prediction.
32
00:01:05,966 --> 00:01:07,600
It's like making up a song for a baby
33
00:01:07,600 --> 00:01:08,800
that combines a known
34
00:01:08,800 --> 00:01:10,500
melody with different words.
35
00:01:17,433 --> 00:01:18,033
To a baby
36
00:01:18,033 --> 00:01:19,266
it sounds totally novel
37
00:01:19,266 --> 00:01:20,866
because they don't know most words
38
00:01:20,866 --> 00:01:22,333
or most melodies.
39
00:01:22,333 --> 00:01:24,133
In the case of generative AI,
40
00:01:24,133 --> 00:01:26,500
generative AI knows billions of melodies
41
00:01:26,500 --> 00:01:27,966
and millions of words.
42
00:01:27,966 --> 00:01:30,966
so what it produces sounds novel to us.
43
00:01:31,566 --> 00:01:33,333
Generative AI models are trained on
44
00:01:33,333 --> 00:01:34,900
massive datasets,
45
00:01:34,900 --> 00:01:36,866
much of which comes from books, articles,
46
00:01:36,866 --> 00:01:39,366
art and media created by real people
47
00:01:39,366 --> 00:01:41,200
available on the internet.
48
00:01:41,200 --> 00:01:42,233
That means that the outputs
49
00:01:42,233 --> 00:01:42,966
might remix
50
00:01:42,966 --> 00:01:44,633
or resemble copyrighted works
51
00:01:44,633 --> 00:01:46,433
without proper attribution.
52
00:01:46,433 --> 00:01:48,266
We need to help students understand.
53
00:01:48,266 --> 00:01:50,300
Just because AI generates content
54
00:01:50,300 --> 00:01:51,633
doesn't mean they own it
55
00:01:51,633 --> 00:01:53,133
or can use it freely.
56
00:01:53,133 --> 00:01:54,500
This raises important lessons
57
00:01:54,500 --> 00:01:55,800
about digital citizenship
58
00:01:55,800 --> 00:01:58,466
and respecting creators.
59
00:01:58,466 --> 00:02:00,100
Students can use AI to generate
60
00:02:00,100 --> 00:02:02,033
essays, homework, or artwork
61
00:02:02,033 --> 00:02:03,700
that looks like their own work,
62
00:02:03,700 --> 00:02:04,900
but it isn't.
63
00:02:04,900 --> 00:02:05,700
This challenges
64
00:02:05,700 --> 00:02:07,300
how we measure learning.
65
00:02:07,300 --> 00:02:08,033
Since tools
66
00:02:08,033 --> 00:02:09,433
detecting generative AI
67
00:02:09,433 --> 00:02:11,333
plagiarism are unreliable,
68
00:02:11,333 --> 00:02:12,933
and because AI is everywhere
69
00:02:12,933 --> 00:02:14,533
in the workforce as well,
70
00:02:14,533 --> 00:02:17,533
the answer isn't banning AI tools.
71
00:02:17,566 --> 00:02:18,300
The first answer
72
00:02:18,300 --> 00:02:19,433
is having a conversation
73
00:02:19,433 --> 00:02:22,033
with students about how we can use AI.
74
00:02:22,033 --> 00:02:23,633
When is AI support allowed
75
00:02:23,633 --> 00:02:26,233
like brainstorming, drafting, outlining?
76
00:02:26,233 --> 00:02:28,166
and when is it crossing into plagiarism
77
00:02:28,166 --> 00:02:29,366
like copy pasting
78
00:02:29,366 --> 00:02:31,300
without engagement or learning?
79
00:02:31,300 --> 00:02:32,133
And secondly,
80
00:02:32,133 --> 00:02:34,066
we need to redesign our assignments
81
00:02:34,066 --> 00:02:35,066
to allow for thoughtful
82
00:02:35,066 --> 00:02:36,900
attribution of generative AI
83
00:02:36,900 --> 00:02:37,866
that allows students
84
00:02:37,866 --> 00:02:38,933
to demonstrate their own
85
00:02:38,933 --> 00:02:41,166
skills and understanding.
86
00:02:41,166 --> 00:02:42,866
Since generative AI creates
87
00:02:42,866 --> 00:02:44,933
content by remixing patterns,
88
00:02:44,933 --> 00:02:47,633
it doesn't produce truly original ideas.
89
00:02:47,633 --> 00:02:49,200
It simulates them.
90
00:02:49,200 --> 00:02:50,133
That means if students
91
00:02:50,133 --> 00:02:51,666
lean too heavily on AI,
92
00:02:51,666 --> 00:02:53,133
their own voices, perspectives,
93
00:02:53,133 --> 00:02:55,400
and creativity can get lost.
94
00:02:55,400 --> 00:02:57,866
We play a key role in encouraging balance
95
00:02:57,866 --> 00:03:00,400
using AI as a partner to spark ideas
96
00:03:00,400 --> 00:03:02,100
and not as a replacement for students
97
00:03:02,100 --> 00:03:04,400
authentic contributions.
98
00:03:05,066 --> 00:03:06,166
There are four things we
99
00:03:06,166 --> 00:03:07,466
can do to support students
100
00:03:07,466 --> 00:03:09,900
with their usage of generative AI.
101
00:03:09,900 --> 00:03:10,666
One.
102
00:03:10,666 --> 00:03:12,266
Design assignments and assessments
103
00:03:12,266 --> 00:03:14,000
that allow attributed usage
104
00:03:14,000 --> 00:03:15,566
of generative AI.
105
00:03:15,566 --> 00:03:18,066
Instead of trying to catch AI use,
106
00:03:18,066 --> 00:03:20,500
use it as a way to enhance the learning.
107
00:03:20,500 --> 00:03:22,666
Two, make the process of using
108
00:03:22,666 --> 00:03:24,266
generative AI transparent
109
00:03:24,266 --> 00:03:26,633
and teach good practices to students.
110
00:03:26,633 --> 00:03:27,566
Co-create prompts
111
00:03:27,566 --> 00:03:29,600
and evaluate results together.
112
00:03:29,600 --> 00:03:30,333
Three.
113
00:03:30,333 --> 00:03:32,566
Discuss authorship and integrity.
114
00:03:32,566 --> 00:03:33,700
When is AI helpful
115
00:03:33,700 --> 00:03:35,266
and when is it damaging?
116
00:03:35,266 --> 00:03:36,233
And four,
117
00:03:36,233 --> 00:03:38,700
discuss overreliance on AI.
118
00:03:38,700 --> 00:03:39,866
Because it could be a convincing
119
00:03:39,866 --> 00:03:40,800
thought partner,
120
00:03:40,800 --> 00:03:41,966
it's important to discuss
121
00:03:41,966 --> 00:03:42,766
habits of mind
122
00:03:42,766 --> 00:03:44,100
that help students retain
123
00:03:44,100 --> 00:03:46,200
their self-governance.
124
00:03:46,200 --> 00:03:48,300
Generative AI doesn't think
125
00:03:48,300 --> 00:03:49,500
it remixes patterns
126
00:03:49,500 --> 00:03:51,733
into something that feels new.
127
00:03:51,733 --> 00:03:52,533
The power comes
128
00:03:52,533 --> 00:03:54,233
not from what it generates,
129
00:03:54,233 --> 00:03:54,866
but from how
130
00:03:54,866 --> 00:03:55,766
we use the outputs
131
00:03:55,766 --> 00:03:58,200
in meaningful and ethical ways.
132
00:03:58,200 --> 00:03:59,433
As educators,
133
00:03:59,433 --> 00:04:01,200
knowing how generative AI works
134
00:04:01,200 --> 00:04:02,066
helps us separate
135
00:04:02,066 --> 00:04:04,266
the hype from the reality
136
00:04:04,266 --> 00:04:05,333
so we can make wise
137
00:04:05,333 --> 00:04:06,766
choices for our classrooms.
