Martin Luther King and the Ghost in the Machine

1
Martin Luther King and the Ghost in the Machine

• Kalamazoo College 2003
• MLK Week Teach-in

2
The birth of the modern US Civil Rights Movement

• Dec. 1 1955 Rosa Parks refuses to move to the
  back of the bus.
• Bus boycott organized King elected president of
  Montgomery Improvement Association.
• Nov. 13, 1956 US Supreme Court declares race
  segregation illegal boycott ends, and full,
  integrated service restored.
• King achieves national prominence as civil rights
  leader.

3
The birth of artificial intelligence research

• Summer, 1956 First Darthmouth Conference on
  AIWe propose that a 2 month, 10 man study of
  artificial intelligence be carried out during the
  summer of 1956 at Dartmouth College in Hanover,
  New Hampshire.
• First common use of the term artificial
  intelligence
• http//www-formal.stanford.edu/jmc/history/dartmou
  th/dartmouth.html

4
1956 Dartmouth AI Conference Attendees
5
Lost opportunties

• The Civil Rights movement and AI research started
  at the same time,
• They never talked to each other.
• What opportunities were lost as a result?

6
King on Technology

• Warns against technology as Moloch
• Worried about automation leading to people
  being thrown out of work.
• Worried about destructive power of violence.

7
Opportunities to be of (peaceful) service to the
community

• Majority of AI research has always been funded by
  the military.
• But AI problems are everywhere (although not
  always fundable).
• The opportunity to found an non-racist, inclusive
  science.

8
King on technology

• Automation can be used to generate an abundance
  of wealth for people. Our society, with its
  ability to perform miracles with machinery has
  the capacity to make miracles for men--if it
  values men as highly as it values machines (From
  If the Negro wins, Labor wins.)
• Through our scientific and technological genius,
  we have made of this world a neighborhood and yet
  we have not had the ethical commitment to make of
  it a brotherhood. (From Remaining awake through
  a great revolution)

9
The Ghost in the Machine

• Gilbert Ryle description of Cartesian dualism
• Body is different from Mind/Soul, etc. (see your
  philosophy teacher)

10
Teaching Whiteness

• Teaching Whiteness, The End of Innocence, Gail B.
  Griffin, to be published
• Stories and reflections and critical essays on
  of being White, teaching writing at Kalamazoo
  College.
• You did go to chapel today, didnt you?

11
A quote from Teaching Whiteness
The irony (or paradox, or both) of whiteness is
that its failure to name itself, while it
arrogates one kind of godlike power (the power of
universality and ubiquity), denies another. For
to be universal and ubiquitous--to be Everything,
Everywhere--is in fact to be Nothing, and
Nowhere, in particular.As the absent agent in a
passive construction, whiteness erases itself.
White language says, in short, I am not here I
do not exist. It does so, of course, to avoid
implicating itself in the relations, past and
present, of racism. But the price for such
exoneration is eternal absence,
non-being--ghostliness.
12
Another missed opportunity

• To reflect on, model, build demos about AIs own
  Whiteness.

13
What is AI?

• Dartmouth goals
• every aspect of learning or any other feature of
  intelligence can in principle be so precisely
  described that a machine can be made to simulate
  it.
• Automatic computers, use language, neuron nets,
  theory of computation, self-improvement,
  abstractions, randomness and creativity

14
What is AI?
15
AI is hard

• Language is AI-Hard
• Vision is AI-Hard
• Planning is AI-Hard
• . Is AI-Hard
• Schanks medieval view of AI

16
What if there are only so many ideas to discover?
Erdöss Gods book of Proofs
Keplers Thinking Gods Thoughts After Him
17
Idea discovery
A researcher is a kind of an experiment the
probability that a given researcher will discover
an idea is P(n,m).
Lets assume ideas are independent researchers
are independent P(n,m) is constant, call it p.
18
Idea discovery (2)
The probability that an idea is discovered by at
least one researcher 1-(1-p)m
19
Idea discovery (3)
How to improve success? 1-(1-p)m
Increase the exponential. 1-(1-p)m
20
Youre handpicking the invite list to
Dartmouthup to 100!
All US
White, non-Hispanic US
White, non-Hispanic Males in US
p.01
.643
100
.587
88
.351
43
1940
.634
100
.500
69
.289
34
2000
21
Idea discovery (model 2)
N good ideas
M researchers
Ideas and people are not colorless.
Lets assume ideas are independent researchers
are independent but P(n,m) is greater if the
color of n and m is the same than if they are
different.
22
Idea discovery (model 2)
N good ideas
M researchers
nb number of blue ideas ng number of green
ideasmb no. of blue researchers mg green
researchersp probability if colors match p?
prob. If not match
Let I(p,m)1-(1-p)m with colors, average
discoveries are
I(nb,p,mb)(1-I(nb,p,mb)) I(nb,p?,mg) for
blue ideas I(ng,p,mg)(1-I(nb,p,mb))
I(ng,p?,mb) for green ideas
23
Back to Dartmouth
White, non-Hispanic Males Green All others Blue
White, non-Hispanic Males Green WNH Females Blue
White, non-Hispanic Males in US only
p .01p? .001
.423
43,57
.385
43,45
.197
43,0
1940
.419
34,66
.317
34,35
.161
34,0
2000
24
And if color of ideas reflects idealized color
of researcher
White, non-Hispanic Males Green All others Blue
White, non-Hispanic Males Green WNH Females Blue
White, non-Hispanic Males in US only
p .01p? .001
.428
43,57
.386
43,45
.175
43,0
1940
.445
34,66
.318
34,35
.121
34,0
2000
25
A conclusion

• These green ideas, which we (AI researchers)
  thought were colorless, lie dormant. I suspect
  they are angry.
• I suspect these colorless green ideas sleep
  furiously.

26
What are some of these ideas?

• Themes of being human which are not captured by
  viewing the field (artificial intelligence) as
  building agents that engage in rational action.

27
Dr. King is my research advisor

• Justice, mercy, conversion, forgiveness,
  violence, revenge, race, politics, resistance,
  persuasion, honor, dignity, sacrifice, love, evil
  
• (To be fair, some researchers have done AI
  research, especially when doing story
  understanding)

28
Analogical reasoning

• It is always difficult to get out of Egypt, for
  the Red Sea always stands before you with
  discouraging dimensions. And even after youve
  crossed the Red Sea, you have to move through a
  wilderness with prodigious hilltops of evil and
  gigantic mountains of opposition. (Kings sermon
  during boycott)

29
Missed Opportunities

• To be of peaceful, just service.
• To found an anti-racist science.
• whose practitioners reflected the makeup of
  society,
• and came to a better understanding of race and
  Whiteness in our cognitive models.
• To model and demonstrate the fuller strands of
  what it means to be human.

30
What could AI be?
Thinking rationally.
Thinking like a human.
Acting rationally.
Acting like a human.
Acting humanely. (humane computing cognitive
technology)
Being like a human. (computational humanism)