-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathAI.html
More file actions
executable file
·183 lines (174 loc) · 9.07 KB
/
Copy pathAI.html
File metadata and controls
executable file
·183 lines (174 loc) · 9.07 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<title>home.html</title>
<meta http-equiv="Content-Type" content="text/html;charset=utf-8"/>
<link rel='stylesheet' href='./simple.css'>
<link rel='stylesheet' href='./my-custom.css'>
</head>
<body>
<head>
<link rel="stylesheet" href="styles.css">
</head>
<h1
id="im-sorry-dave-im-afraid-i-cant-do-that-what-history-could-teach-us-about-designing-humane-ai">I’m
sorry, Dave, I’m Afraid I Can’t Do That: What History Could Teach Us
About Designing Humane AI</h1>
<h3 id="course-leader-tom-naps">Course Leader: Tom Naps</h3>
<p>Welcome to the course home page for the spring 2024 OLLI course
“<em>I’m sorry, Dave, I’m Afraid I Can’t Do That: What History Could
Teach Us About Designing Humane AI</em>”. The course will meet in
Magrath Library Room 8 for seven Thursdays 12:30 - 2:00 starting March
28. Feel free to contact me (<a href="mailto:thomas.naps@gmail.com"
class="uri">mailto:thomas.naps@gmail.com</a>) anytime you have questions
or concerns about the course.</p>
<p><strong>Course description from OLLI catalog:</strong> When HAL
refused Dave’s order to open the pod bay doors in Stanley Kubrick’s
movie adaptation of Arthur C. Clarke’s 2001: A Space Odyssey, was it
going rogue or making a rational choice among options it had?
Philosophers, writers, and scientists have long speculated on thinking
machines. Have we now reached an inflection point in the relationship we
will have with AI going forward? What could it be capable of? What do we
want it to be capable of? How can we align the potential of AI with our
own weaknesses and strengths? We will explore the answers to those
questions and more.</p>
<hr />
<h2 id="course-outline">Course outline</h2>
<ul>
<li>Week 1 - AI, please do the right thing! Concerns lying ahead as we
emerge from the AI winter. ( <a href="Week1.pptx">Powerpoint version of
slides</a> ) ( <a href="Week1.pdf">PDF version</a> )</li>
<li>Week 2 - Moravec’s paradox. Big data meets video game hardware and
AI’s trajectory changes from rule-based heuristics to deep learning. (
<a href="Week2.pptx">Powerpoint version of slides</a> ) ( <a
href="Week2.pdf">PDF version</a> )</li>
<li>Week 3 - What can we learn from three case studies? IBM’s Watson,
driverless cars, and facial recognition. ( <a
href="Week3.pptx">Powerpoint version of slides</a> ) ( <a
href="Week3.pdf">PDF version</a> )</li>
<li>Week 4 - LLMs: Neural nets on steroids. ( <a
href="Week4.pptx">Powerpoint version of slides</a> ) ( <a
href="Week4.pdf">PDF version</a> )</li>
<li>Week 5 - Framing the issues and open questions that your small
groups will focus on next week. ( <a href="Week5.pptx">Powerpoint
version of slides</a> ) ( <a href="Week5.pdf">PDF version</a> ). The
issues are:
<ol type="1">
<li>Privacy</li>
<li>Monopolization</li>
<li>Geopolitics and national security</li>
<li>Monetization</li>
<li>Free speech</li>
<li>Intellectual property</li>
<li>Bias in training data</li>
<li>Human work in the context of AI</li>
<li>Deep fakes</li>
</ol></li>
<li>Week 6 - Work in small discussion groups, chosen on the basis of
interest in a common topic from Week 5. Tom will circulate, trying to
provide “assistance” and offering the opportunity to run a query your
group may have on Perplexity.ai</li>
<li>Week 7 - Informal group presentations using a “choose your
adventure” format in which your group identifies options we have in
deciding how to “control” AI on your issue and then predicting the
outcome that will result when a particular option is chosen. Here are
the groups that emerged during Week 6:
</ul>
<p>Slides for each week will be posted shortly before class.</p>
<hr />
<h2 id="additional-resources-for-deeper-exploration">Additional
resources for deeper exploration</h2>
<ul>
<li><a
href="https://www.goodreads.com/book/show/144405196-the-worlds-i-see"><em>The
Worlds I See: Curiosity, Exploration, and Discovery at the Dawn of
AI</em> by Fei-fei Li</a>. Li is one of the world’s leading researchers
in image recognition. This book interweaves the history of this area of
AI with a memoir detailing Li’s journey from a young girl in China’s
mail-centered education system to her present positions as Co-Director
of Stanford’s Institute for Human-Centered Artificial Intelligence and
co-founder/board chair of the <a href="https://ai-4-all.org/">AI4ALL
organization</a>, a non-profit dedicated to improving AI by increasing
the diversity of those involved in it and thereby correcting <em>the
alignment problem</em> (see next resource).</li>
<li><a
href="https://www.goodreads.com/book/show/50489349-the-alignment-problem"><em>The
Alignment Problem: Machine Learning and Human Values</em> by Brian
Christian</a>.</li>
<li><a
href="https://www.goodreads.com/book/show/136342721-unmasking-ai"><em>Unmaksing
AI: My Mission to Protect What Is Human in a World of Machines</em> by
Joy Buolamwini</a>. When she was an MIT graduate student embarking on a
research project in facial recognition, Buolamwini discovered that
nearly all existing facial recognition systems failed miserably when it
came to subjects who were, like her, female and black. That inspired her
to dedicate her career to addressing issues of racial and gender bias in
AI systems.</li>
<li><a href="EmilyBender-You-Are-Not-a-Parrot-in-NTMag.pdf"><em>You Are
Not a Parrot</em> by Emily Bender</a>. Bender is a computational
linguist at U of Washington. This NYMagazine article presents a shorter
layperson perspective on a classic academic paper by Bender in which she
detailed differences between understanding and merely being able to
mimic expression in a natural language. A Youtube video discussion of
her research in this area appears <a
href="https://www.turing.ac.uk/events/dangers-stochastic-parrots">here</a>.</li>
<li><a
href="https://www.goodreads.com/book/show/90590134-the-coming-wave"><em>The
Coming Wave: Technology, Power, and the Twenty-first Century’s Greatest
Dilemma</em> by Mustafa Suleyman</a>. Suleyman was a co-founder of
Google’s DeepMind, then started up his own company <em>Inflection
AI</em>, and was recently hired by Microsoft (along with much of
Inflection’s staff) to lead their consumer AI business. In this book, he
gives his perspective on a future where we will need to make careful
choices to avoid an AI dystopia.</li>
<li><a
href="https://www.goodreads.com/book/show/62294487-the-little-learner"><em>The
Little Learner: A Straight Line to Deep Learning</em> by Daniel P.
Friedman and Anurag Mendhekar</a>. A good book if you’re interested in a
deeper dive into the mathematics behind neural nets. Written in a
uniquely conversational style but still requires a slow reading and
ideally tinkering with the book’s code that you can download on your
computer.</li>
<li><a
href="https://www.goodreads.com/book/show/75293505-your-face-belongs-to-us"><em>Your
Face Belongs to Us: A Secretive Startup’s Quest to End Privacy as We
Know It</em> by Kashmir Hill</a>. Hill is the privacy reporter for the
NY Times. This book is a fascinating account of how <em>ClearviewAI</em>
went full-speed ahead on a project that both Google and Facebook deemed
too risky to continue.</li>
<li><a
href="https://www.goodreads.com/book/show/56097578-god-human-animal-machine"><em>God,
Human, Animal, Machine: Technology, Metaphor, and the Search for
Meaning</em> by Meghan O’ Gieblyn</a>. The author, raised in an
evangelical family, reflects on how she has changed and what being human
might mean in an age of AI.</li>
<li><a
href="https://www.goodreads.com/book/show/195791688-burn-book"><em>Burn
Book: A Tech Love Story</em> by Kara Swisher</a> Tech’s leading critical
journalist reflect on her experience interacting with the leaders of
technology, many of whom are deeply involved in how AI will factor into
our future. She’s not easy on them when it comes to their motives.
Earlier in March, as part of her book tour, she spoke at the Westminster
Forum in Minneapolis - you can view a <a
href="https://www.youtube.com/watch?v=ybKtsfydqY4">recording of the
event here</a>.</li>
<li><a
href="https://www.goodreads.com/book/show/61089453-four-battlegrounds"><em>Four
Battlegrounds: Power in the Age of Artificial Intelligence</em> by Paul
Scharre</a>. A geopolitical perspective on AI.</li>
<li><a
href="https://www.goodreads.com/book/show/60623908-power-and-prediction"><em>Power
And Prediction: The Disruptive Economics of Artificial Intelligence</em>
by Ajay Agrawal, Joshua Gans, and Avi Goldfarb</a> If our minds are
predictive machines, how well can AI predict and influence what those
minds are going to predict?</li>
<li><a
href="https://www.goodreads.com/book/show/60784561-the-battle-for-your-brain"><em>The
Battle for Your Brain: Defending the Right to Think Freely in the Age of
Neurotechnology</em> by Nita A. Farahany</a> AI wants to learn how to
“be one” with our human brains. Do we want that?</li>
</ul>
</body>
</html>