Article
Giving datasets a factsheet
A proposal for documenting where a dataset comes from, what it contains and what it is intended for.
Why read it?
Sets out questions to ask about data before using it.
Keep in mind
Completing a description does not guarantee that the data are representative or suitable for a particular purpose.
Reference details
Title as recorded: Datasheets for datasets
Author(s): Timnit Gebru; Jamie Morgenstern; Briana Vecchione; Jennifer Wortman Vaughan; Hanna Wallach; Hal Daumé III; Kate Crawford · Year: 2021
Publication / host: Association for Computing Machinery (ACM)
Source language: English
DOI: 10.1145/3458723
KP-R-12B69AB3DCFF
Classic text
Machiavelli on power and government
A historical English translation of Il Principe, discussing power, means and credibility.
Why read it?
Offers a historical starting point for questions about power, ends and the means used to pursue them.
Keep in mind
Read the passages in their political and historical context. Machiavelli writes from a particular perspective on power and government.
Reference details
Title as recorded: Machiavelli: The Prince
Author(s): Not recorded · Year: Not recorded
Publication / host: www.gutenberg.org
Source language: English
KP-R-12F7CB9D51EB
Article
Thomson and the trolley problem
Thomson uses the trolley problem to examine moral differences between killing and letting someone die.
Why read it?
An original source for asking why seemingly similar choices can feel different.
Keep in mind
The thought experiments simplify the situation. They are not ready-made decision rules for real emergencies.
Reference details
Title as recorded: Killing, Letting Die, and the Trolley Problem
Author(s): Judith Jarvis Thomson; The Hegeler Institute · Year: 1976
Publication / host: Oxford University Press (OUP)
Source language: English
DOI: 10.5840/monist197659224
KP-R-13DE830B701B
Preprint
Possible misuse of AI
Brundage and colleagues describe ways AI could be used maliciously and possible preventive measures.
Why read it?
Provides concrete scenarios for discussing human misuse of AI.
Keep in mind
The report dates from 2018. Its scenarios describe possibilities, not evidence that all of them will occur.
Reference details
Title as recorded: The Malicious Use of Artificial Intelligence: Forecasting, Prevention, and Mitigation
Author(s): Brundage, Miles; Avin, Shahar; Clark, Jack; Toner, Helen; Eckersley, Peter; Garfinkel, Ben; Dafoe, Allan; Scharre, Paul; Zeitzoff, Thomas; Filar, Bobby; Anderson, Hyrum; Roff, Heather; Allen, Gregory C.; Steinhardt, Jacob; Flynn, Carrick; hÉigeartaigh, Seán Ó; Beard, SJ; Belfield, Haydn; Farquhar, Sebastian; Lyle, Clare; Crootof, Rebecca; Evans, Owain; Page, Michael; Bryson, Joanna; Yampolskiy, Roman; Amodei, Dario · Year: Not recorded
Publication / host: arxiv.org
Source language: English
KP-R-14E0D91B4194
ArticleStarting point
The ironies of automation
Bainbridge shows why automation can place substantial demands on human expertise precisely when exceptional situations arise.
Why read it?
Explores what people need to remain able to do when systems take over routine work.
Keep in mind
The analysis was published in 1983. Applying it to generative AI also requires examining differences from those earlier systems.
Reference details
Title as recorded: Ironies of automation
Author(s): Lisanne Bainbridge · Year: 1983
Publication / host: Elsevier BV
Source language: English
DOI: 10.1016/0005-1098(83)90046-8
KP-R-159A6572D202
Book
The psychology of information analysis
Heuer describes patterns and limitations in thinking when assessing incomplete information.
Why read it?
Offers ways to examine our own assumptions when evaluating information.
Keep in mind
The book was written for intelligence analysis. Its background matters when applying it elsewhere.
Reference details
Title as recorded: De psychologie van informatieanalyse
Author(s): Not recorded · Year: Not recorded
Publication / host: www.cia.gov
Source language: English
KP-R-19681A308C3C
Article
What does “I'm not sure” do to our trust?
This study examines how people respond when a language model expresses uncertainty.
Why read it?
Encourages us to assess cautious-sounding AI answers on their substance.
Keep in mind
Cautious wording does not necessarily mean that the system accurately assesses its own reliability.
Reference details
Title as recorded: “I’m Not Sure, But…”: Examining the Impact of Large Language Models’ Uncertainty Expression on User Reliance and Trust
Author(s): Sunnie S. Y. Kim; Q. Vera Liao; Mihaela Vorvoreanu; Stephanie Ballard; Jennifer Wortman Vaughan · Year: 2024
Publication / host: www.microsoft.com
Source language: English
KP-R-1BEB6E9F2866
Article
When explanations of AI advice help
A study of circumstances in which explanations accompanying AI advice reduce overreliance.
Why read it?
Explores which explanations help people assess AI advice for themselves.
Keep in mind
The effect depends on the task and on the effort the user makes to evaluate the explanation.
Reference details
Title as recorded: Explanations Can Reduce Overreliance on AI Systems During Decision-Making
Author(s): Helena Vasconcelos; Matthew Jörke; Madeleine Grunde-McLaughlin; Tobias Gerstenberg; Michael S. Bernstein; Ranjay Krishna · Year: 2023
Publication / host: Association for Computing Machinery (ACM)
Source language: English
DOI: 10.1145/3579605
KP-R-1C3739E50FD1
ChapterCounterpoint
Is “no true Scotsman” always a fallacy?
Anderson examines when someone changes the meaning of a category to evade criticism.
Why read it?
Helps assess shifting definitions carefully.
Keep in mind
The particular argument matters. Naming a fallacy is not enough to establish that it has occurred.
Reference details
Title as recorded: Is Flew’s No True Scotsman Fallacy a True Fallacy? A Contextual Analysis
Author(s): Robert Ian Anderson · Year: 2017
Publication / host: Springer International Publishing
Source language: English
DOI: 10.1007/978-3-319-57837-8_19
KP-R-1D3D71EEA730
Article
Banach–Tarski: doubling a ball?
The original mathematical text on dividing and rearranging sets of points.
Why read it?
An original source for a paradox that challenges everyday ideas about space and volume.
Keep in mind
The proof concerns abstract sets and uses the axiom of choice. It cannot be used to double a physical ball of matter.
Reference details
Title as recorded: Sur la décomposition des ensembles de points en parties respectivement congruentes
Author(s): Stefan Banach; Alfred Tarski · Year: 1924
Publication / host: Institute of Mathematics, Polish Academy of Sciences
Source language: French
DOI: 10.4064/fm-6-1-244-277
KP-R-1E827DD630EC
Article
Thinking you knew it all along: does age matter?
This meta-analysis brings together research on age differences in hindsight bias: the feeling, after an outcome, that we saw it coming.
Why read it?
Explores how knowing the outcome can colour memory.
Keep in mind
Average differences between age groups do not tell us how an individual reasons.
Reference details
Title as recorded: Age differences in hindsight bias: A meta-analysis.
Author(s): Julia Groß; Thorsten Pachur · Year: 2019
Publication / host: American Psychological Association (APA)
Source language: English
DOI: 10.1037/pag0000329
KP-R-1EBCD4D9FFF7
Article
Who checks AI within an organisation?
A proposal for auditing algorithmic systems within an organisation, from development through use.
Why read it?
Helps translate accountability for AI into concrete checks.
Keep in mind
These checks require access to information, expertise and follow-up. A documented approach alone is not enough.
Reference details
Title as recorded: Closing the AI accountability gap
Author(s): Inioluwa Deborah Raji; Andrew Smart; Rebecca N. White; Margaret Mitchell; Timnit Gebru; Ben Hutchinson; Jamila Smith-Loud; Daniel Theron; Parker Barnes · Year: 2020
Publication / host: ACM
Source language: English
DOI: 10.1145/3351095.3372873
KP-R-1F8EAF79A9F5
Article
Probability errors in legal fact-finding
Dahlman organises common errors in using probability to assess evidence.
Why read it?
Shows why a convincing-sounding probability calculation can still mislead.
Keep in mind
The article discusses forms of reasoning. It does not advise on an individual case.
Reference details
Title as recorded: A systematic account of probabilistic fallacies in legal fact-finding
Author(s): Christian Dahlman · Year: 2024
Publication / host: SAGE Publications
Source language: English
DOI: 10.1177/13657127231209019
KP-R-207420A9BD6F
Article
Can thinking extend beyond our heads?
Clark and Chalmers ask whether external tools can form part of our thinking.
Why read it?
Explores where our own thinking ends and a tool begins.
Keep in mind
This is a philosophical argument. It does not imply that every tool automatically becomes part of the mind.
Reference details
Title as recorded: The Extended Mind
Author(s): A. Clark; D. Chalmers · Year: 1998
Publication / host: Oxford University Press (OUP)
Source language: English
DOI: 10.1093/analys/58.1.7
KP-R-21EA3BF21B7D
Legislation
The GDPR and automated decisions
The official Dutch text of the GDPR, including Article 22 on decisions based solely on automated processing.
Why read it?
A primary legal source for questions about automated decision-making.
Keep in mind
Read the conditions and exceptions together. A single article does not necessarily establish what applies in a particular situation.
Reference details
Title as recorded: De AVG en geautomatiseerde besluiten
Author(s): Not recorded · Year: Not recorded
Publication / host: eur-lex.europa.eu
Source language: Dutch
KP-R-22151E3C0270
Preprint
Searching where the light falls
Research into how available data influence what researchers and organisations discover.
Why read it?
Invites us to ask what we miss when we study only what is easy to measure.
Keep in mind
This is a preliminary research version. Its conclusions may change in a later version.
Reference details
Title as recorded: The Streetlight Effect in Data-Driven Exploration
Author(s): Johannes Hoelzemann; Gustavo Manso; Abhishek Nagaraj; Matteo Tranchero · Year: 2024
Publication / host: National Bureau of Economic Research
Source language: English
DOI: 10.3386/w32401
KP-R-22CB1C1B6EE0
Classic text
Nietzsche: Ecce Homo
Nietzsche reflects on his work and approach to life, including ideas related to amor fati: affirming one's fate.
Why read it?
Provides background for questions about our relationship with our own lives and past.
Keep in mind
This is a literary and philosophical work, not a tested therapeutic approach.
Reference details
Title as recorded: Nietzsche: Ecce Homo
Author(s): Not recorded · Year: Not recorded
Publication / host: www.gutenberg.org
Source language: English
KP-R-230138C6408D
Article
What students think others find normal
Prentice and Miller examine the gap between students' own views on drinking and the views they attribute to their group.
Why read it?
Shows how people may adapt to a group norm they have misjudged.
Keep in mind
The research concerns a particular campus setting. It cannot simply be used to describe another group.
Reference details
Title as recorded: Pluralistic ignorance and alcohol use on campus: Some consequences of misperceiving the social norm.
Author(s): Deborah A. Prentice; Dale T. Miller · Year: 1993
Publication / host: American Psychological Association (APA)
Source language: English
DOI: 10.1037/0022-3514.64.2.243
KP-R-2318B2135884
No sources match these filters. Try fewer filters or a different search term.