Jeroen van den Hoven. Omgaan met ethische vraagstukken over digitalisering van het onderwijs.

1. Digital Ethics and
Education
Jeroen van den Hoven
University Professor
Delft University of Technology

2. Contents
1. Introduction: A changing digital world
2. Robotics:
The Future of Work
3. AI and Big Data
Transparency, Privacy, Trust
4. Social media:
Evil on-line

3. 1. Introduction

4. New
AI
Deep Learning
Robotics
Big Data
Quantum Computing
Cloud Computing
Block chain
Internet of Things
Social media and On-
line platforms
3D Printing
5G
Privacy
Accountability
Democracy
Property
Control
Safety
Security
Sustainability
Human dignity
Identity
Social cohesion

5. Confusion
Cyber
Net
E-
Artificial
Virtual
Internet
Digital
Online
Blockchain
Community
Friendship
Person
Intelligence
Property
Crime
Democracy
Education
Organization

6. Vacuum
Blockchain
organisation
Conceptual Vacuum
Policy Vacuum
Design Vacuum

7. Towards a Good Digital Society?
Equality
Liberty
Solidarity
Property
Transparency
Accountability
Democracy
Justice
Privacy
Dignity

9. The Silicon Valley Model

10. Geopolitics and Digital Power

12. Citizen Scores

16. Eric Schmidt CEO Google
have been, more or less what you

17. Google: Alphabet
60.000 employees, annual turnover 70 miljard
Ad revenues
Adwords
Adsense
Google Car
Googledocs
Google glass
Youtube
Android
Gmail
Calendar
Google maps
Google Play
Google hangouts
Google Analytics
Google drive
Double Click (*attributes captured per visit > 100)

18. We want to connect people!
Really?

19. Zuboff: Surveillance Capitalism

21. Harari: datism

22. Rushkoff

24. CASS SUNSTEIN
CHOICE
ARCHITECTURES

25. nudges

26. Persuasive Technology:
Speedbump

27. Persuasive Technology: Dutch
Design

29. CASS R. SUNSTEIN WHY NUDGE?
companies that
provide clear, simple
products would do
poorly in the
marketplace,
because they are not
taking advantage of
people's propensity
to blunder" (p. 11)

30. A/B testing

31. Cambridge Analytica: Nix

33. IBM: 47 psychographic parameters
Includes values
and needs

34. Big Data Machinery

35. Our cognitive deficiencies

36. Exploited: Advanced Behavioral
Science + Big Data + AI

37. Microtargetting

38. Digital Society of having Control or
being Controled

39. Tristan Harris: ex googler

41. Educational System
Nudging
Scoring
Labelling
Testing
Accrediting
Certifying
Measuring
Quantifying
Data collection
$$$$$$$$$$$$$$$$$

42. 2. Robots and the Future
of Work

45. Second Machine Age
Technical Unemployment

46. "Software substitution, whether it's for drivers
... Technology over time will reduce demand
for jobs, particularly at the lower end of skill
set. ... 20 years from now, labor demand for
lots of skill sets will be substantially lower.

55. The Future of Work

56. 3. Artificial Intelligence

58. in computing power, availability of data and
progress in algorithms have turned AI into one of the
most strategic technologies of the 21st century. The
stakes could not be higher. The way we approach AI
will define the world we live in
to the European Parliament, the European Council, The Council,
the European Economic and Social Committee and the
April 2018) {COM (2018) 237 Final}, 13-14
European Commission

59. Science and
Technology required
to make systems
behave as in they
do in Movies
AI

60. Neural Nets

61. Neural nets
Deep Learning

62. Deep neural nets

63. Is this Thinking? Understanding?

64. John Searle

65. Edgser Dijkstra

66. Powerful tools and computational
machinery:
Algorithms sift through Big Data and find
Classify, categorize, label, characterize,
predict, decide, recognize, sense,
distinguish, represent, choose, target,
judge, sentence, refuse, nudge, coach,
profile, select,

68. Go World Champion

74. Three Problems
1.Transparency
2. Privacy
3.Trust

75. Some of our most powerful
technologies are the most
intransparent ones
1. Transparency: Deep
Learning is glorified
correlation

76. Spurious Correlations

79. Machine Learning: Black Box

80. Justification and explainability

82. Explainability : GDPR (2018)
if an algorithm is used on EU citizens one must publish
the following:
a. enough technical details around the model selection
and training process including the origination and type
of the data set used for training.
b. understand the importance of the model in public
the impact of false positives
and false negatives and document them and
c. set up systems to educate the data-subject on why
not to opt-out of the model- prediction process.

83. Algorithmic citizenship

84. FDA Approves Crisis-Predicting Algorithm to
Save Hospital Patients From Early Death
Predicting death

85. Aspire Health, USA
Cost reduction in palliative care

86. Three Problems
1.Transparency
2. Privacy
3.Trust

87. Big Data

88. Big Data

89. Convergence digital technology
Cloud Computing
Social Networking
Mobile Computing
Robotics
Augmented Reality
Sensors
Internet of Things

91. Too much information
2.5×1018 bytes

92. shopping
Preferences, emotions
Social relations
Location, movement

93. Twitter map

99. Three Problems
1.Transparency
2. Privacy
3.Trust

102. Open Science

103. 4. Social Media

106. Cyberbullying

107. Revenge Porn

108. Pedophilia

109. Catfishing

110. Public shaming

111. Self harm

112. Self harm

113. Anorexia/orthorexia

114. Radicalization
Zac Chesser
Dylan
Roof

115. Prankster industry and
social media

117. How is the Online world
different?
Some morally relevant features

118. Evil Online

119. Advantages of the Web
Successful Knowledge (epistemic) practice
1. Economy
2. Power
3. Reliability
4. Fecundity
Coordination, Collaboration
Connectivity (communication, interaction, transaction)

120. Evil Online: Moral Fog
1. Selectivity
2 . Homophiliy,
Stigmergy
3. Anonymity
4. Virtuality
5. Voluntariness
6. Positionality
7. Domesticity
8. Isolation
9. Public Private
Blurring
10. Addictiveness
11. Jurisdiction
12. Interpretative
flexibility

121. 1. Selectivity
Algorithmic Bias

122. Algorithmic Bias in Newsfeed

123. FB News feed algorithm

124. Recommender systems

125. Filter Bubbles; Echo chambers
Likes
Recommender Systems
Customization
Daily Me

126. 2. Homophily

127. 2. Stigmergy

128. 2. Information Cascades:

129. 3. Anonymity

130. 4. Virtuality

131. 5. Voluntary identity: Trust

132. 6. Positionality

133. 7. Domesticity

134. 8. Isolation
9. Isolation
10.

135. 9. Public Private confusion

136. 10. Design for Addiction

137. 11. Jurisdiction

138. 12. Interpretative Flexibility

139. Problem of relevant description
There are innumerous true
descriptions of actions and behavior.
Some are more (morally) relevant
than others
People can pick the ones that make
their behavior look innocuous

140. Result:
Moral Fog

141. Implications of the moral Fog
Disorientation in a space of moral salience and relevance
Problem of Relevant Descriptions
An action can be truthfully described in numerous ways
Perpetrators of Evil Online pick out of all descriptions the

142. Evil is thought-defying: a profound dumbness with respect to
the world and the Other, and systematic turning away from
humanness, morality and conscience, in some cases as in
Nazism to the point of reversing the concepts of good and
evil.

143. Normal people turning bad

144. Image of man: A beast
Good laws make good people

145. Image of man: Angel

146. Good and bad guys; normal
distribution

147. A cooperative species

148. Veneer Theory of ethics

149. Pro social behaviour

150. A Digital world that is
designed and built
Can it accommodate morality, empathy and pro-sociality?

151. Churchill

152. Moral Values (both right and
wrong ones, ours and
, get expressed
(intentionally and
inadvertently) in our digital
world

153. L. Winner: Do artefacts have
politics?

154. Design for Addiction

155. The Importance of
Design

156. Blue eyes, brown eyes

157. Situationism

158. Milgram

159. Ash Conformity

160. Stanford Students: prison
experiment

161. Artifacts have politics
Technology is not neutral
Design shapes the life world of people
Design nudge, shape, constrain and
enable
New type of responsibility for government!
Responsibility: Design of situation

162. How to think about it
ethically?

164. Key Problem
21st Century: Value Sensitive Design

165. Identity Management
Systems
Authorization Matrix
Procedures
Regulations
Incentive structures
Auction mechanisms
Voting mechanism
Monitoring and inspection
Interfaces
Infrastructures
Algorithms
Reputation
Filter
Reach
Classification
Categorization
Ontologies
Code
Protocols
Integrity constraints
Architectures
Governance arrangements
Values Built into information
Systems

166. Design for X
Design for privacy
Design for security
Design for inclusion
Design for sustainability
Design for democracy
Design for safety
Design for transparency
Design for accountability
Design for responsibility

167. Design for critical thinking,
reflection and dialogue

170. Three threats digital society:
challenges for educational systems
1. Assault on the autonomous
individual: be prepared
2. Employment and contribution to
society: be prepared
3. Social media, pro-sociality, moral
development: be prepared