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
ArticleStarting point
AI and the illusion of scientific understanding
Messeri and Crockett examine how AI can make us feel that we understand something while blind spots remain.
Why read it?
Explores the difference between a convincing answer and genuine understanding.
Keep in mind
The article develops a perspective on the problem. It does not measure how often or how strongly researchers experience it.
Reference details
Title as recorded: Artificial intelligence and illusions of understanding in scientific research
Author(s): Lisa Messeri; M. J. Crockett · Year: 2024
Publication / host: Springer Science and Business Media LLC
Source language: English
DOI: 10.1038/s41586-024-07146-0
KP-R-3F0F5926316F
ArticleStarting point
Monty Hall starts with the rules
Gill explains why calculating the probabilities requires knowing the rules and the host's behaviour.
Why read it?
Encourages us to define a problem before doing the calculation.
Keep in mind
The familiar two-thirds probability applies only when the relevant assumptions hold.
Reference details
Title as recorded: The Monty Hall problem is not a probability puzzle* (It's a challenge in mathematical modelling)
Author(s): Richard D. Gill · Year: 2011
Publication / host: Wiley
Source language: English
DOI: 10.1111/j.1467-9574.2010.00474.x
KP-R-6CB442F75993
Reference workStarting point
The heap paradox: when does something become a heap?
An overview of the history of the sorites paradox and attempts to deal with vague boundaries.
Why read it?
Explores when many small changes together make a difference.
Keep in mind
Philosophers disagree about a solution. The overview discusses several approaches.
Reference details
Title as recorded: Sorites Paradox
Author(s): Raffman, Diana; Hyde, Dominic · Year: 1997
Publication / host: plato.stanford.edu
Source language: English
KP-R-8034F0F41A65
ArticleStarting point
The original Simpson's paradox
Simpson examines how the association between two characteristics may differ when groups are considered together or separately.
Why read it?
An original source for surprising differences between subgroup and aggregate figures.
Keep in mind
A statistical association does not, by itself, establish a cause.
Reference details
Title as recorded: The Interpretation of Interaction in Contingency Tables
Author(s): E. H. Simpson · Year: 1951
Publication / host: Oxford University Press (OUP)
Source language: English
DOI: 10.1111/j.2517-6161.1951.tb00088.x
KP-R-9BB01B5B90C1
Classic textStarting point
Chesterton's fence: understand before changing
In The Thing, Chesterton recommends finding out why something exists before removing it.
Why read it?
Encourages questions about an arrangement's original purpose when considering change.
Keep in mind
The argument calls for investigating function. It does not require preserving every existing rule.
Reference details
Title as recorded: Chestertons hek: eerst begrijpen, dan veranderen
Author(s): Not recorded · Year: Not recorded
Publication / host: catholiclibrary.org
Source language: English
KP-R-A19B28CF895A
ArticleStarting point
Depending on others for knowledge
Hardwig examines why much of our knowledge depends on other people's research and judgement.
Why read it?
Explores how we can know something without checking everything ourselves.
Keep in mind
The article develops a philosophical argument. Depending on others does not require trusting them blindly.
Reference details
Title as recorded: Epistemic Dependence
Author(s): John Hardwig · Year: 1985
Publication / host: Philosophy Documentation Center
Source language: English
DOI: 10.2307/2026523
KP-R-A6E0CC406091
ArticleStarting point
Mary's room: knowing and experiencing
Jackson asks whether knowing all the physical facts also means knowing what a conscious experience feels like.
Why read it?
An original source for asking whether experience teaches something beyond a description.
Keep in mind
This is a thought experiment, not experimental evidence against a complete physical explanation.
Reference details
Title as recorded: Epiphenomenal Qualia
Author(s): Frank Jackson · Year: 1982
Publication / host: Oxford University Press (OUP)
Source language: English
DOI: 10.2307/2960077
KP-R-A82C11173E3E
BookStarting pointCounterpoint
Ostrom: managing shared resources together
In Governing the Commons, Ostrom investigates how communities organise the use of shared resources.
Why read it?
Offers a counterpoint to the idea that shared resources inevitably become depleted.
Keep in mind
No approach works everywhere. Local rules, monitoring and circumstances matter.
Reference details
Title as recorded: Governing the Commons
Author(s): Elinor Ostrom · Year: 1990
Publication / host: Cambridge University Press
Source language: English
DOI: 10.1017/cbo9780511807763
KP-R-A9AEAD521976
ArticleStarting point
Testing yourself to help remember
Roediger and Karpicke examine how actively retrieving information can help people remember it later.
Why read it?
Provides research for asking why self-testing can be part of learning.
Keep in mind
The result concerns the learning tasks studied. Not every form of testing will have the same effect.
Reference details
Title as recorded: Test-Enhanced Learning
Author(s): Henry L. Roediger; Jeffrey D. Karpicke · Year: 2006
Publication / host: SAGE Publications
Source language: English
DOI: 10.1111/j.1467-9280.2006.01693.x
KP-R-B8BA0A6DDBBB
ArticleStarting point
To trust or to think?
A study examines interventions that prompt people to think before following AI advice.
Why read it?
Provides research on how to create room for independent judgement.
Keep in mind
The results come from particular decision tasks. Their relevance to other working environments needs to be checked.
Reference details
Title as recorded: To Trust or to Think
Author(s): Zana Buçinca; Maja Barbara Malaya; Krzysztof Z. Gajos · Year: 2021
Publication / host: Association for Computing Machinery (ACM)
Source language: English
DOI: 10.1145/3449287
KP-R-CDF0CB7F1BAA
Reference workStarting point
Recognising fallacies without labelling people
An extensive overview of fallacies and the importance of an argument's context.
Why read it?
Provides reference material for careful examination of arguments.
Keep in mind
Assessing an argument requires examining the reasoning, even when you recognise a fallacy's name.
Reference details
Title as recorded: Drogredenen herkennen zonder etikettenplakken
Author(s): Not recorded · Year: Not recorded
Publication / host: iep.utm.edu
Source language: English
KP-R-D7938D0306FA
ArticleStarting point
When the score becomes the goal
Bevan and Hood examine how targets and measurement in English healthcare can steer behaviour towards better figures.
Why read it?
Helps ask whether a better score means a better reality.
Keep in mind
The study concerns a particular organisational setting. Not every performance measure has the same consequences.
Reference details
Title as recorded: WHAT’S MEASURED IS WHAT MATTERS: TARGETS AND GAMING IN THE ENGLISH PUBLIC HEALTH CARE SYSTEM
Author(s): GWYN BEVAN; CHRISTOPHER HOOD · Year: 2006
Publication / host: Wiley
Source language: English
DOI: 10.1111/j.1467-9299.2006.00600.x
KP-R-E2E1A3B7343F
ArticleStarting point
The Abilene paradox
Harvey describes how a group can choose something that its individual members do not actually want.
Why read it?
Helps ask whether agreement reflects a genuinely shared preference.
Keep in mind
The familiar family story illustrates the idea. It is not a controlled experiment.
Reference details
Title as recorded: The abilene paradox: The management of agreement
Author(s): Jerry B. Harvey · Year: 1974
Publication / host: Elsevier BV
Source language: English
DOI: 10.1016/0090-2616(74)90005-9
KP-R-E4C479470E15
ArticleStarting point
Performing well is not the same as learning
This review discusses the difference between immediate performance and learning something that remains available later.
Why read it?
Helps distinguish performance with assistance from what someone can later do independently.
Keep in mind
Investigating that difference requires measurement at different times. Better immediate performance can still be valuable in itself.
Reference details
Title as recorded: Learning Versus Performance
Author(s): Nicholas C. Soderstrom; Robert A. Bjork · Year: 2015
Publication / host: SAGE Publications
Source language: English
DOI: 10.1177/1745691615569000
KP-R-E90C665B1234
Classic textStarting point
Plutarch and the Ship of Theseus
In Theseus 23.1, Plutarch describes the replacement of a ship's parts and the question of whether it remains the same ship.
Why read it?
An original source for a familiar question about identity and change.
Keep in mind
This is a historical account that prompts a philosophical question, not a modern experiment.
Reference details
Title as recorded: Plutarchus en het schip van Theseus
Author(s): Not recorded · Year: Not recorded
Publication / host: penelope.uchicago.edu
Source language: English
KP-R-EBFAACE71A80
ArticleStarting point
AI help with mathematics: what can you do afterwards?
A field experiment examines pupils' performance on mathematics tasks with AI assistance and what they can subsequently do without it.
Why read it?
Helps distinguish good performance with AI from what a pupil has learned independently.
Keep in mind
Read the published correction, DOI 10.1073/pnas.2518204122. The conclusions depend on the setting and form of AI support studied.
Reference details
Title as recorded: Generative AI without guardrails can harm learning: Evidence from high school mathematics
Author(s): Hamsa Bastani; Osbert Bastani; Alp Sungu; Haosen Ge; Özge Kabakcı; Rei Mariman · Year: 2025
Publication / host: National Academy of Sciences
Source language: English
DOI: 10.1073/pnas.2422633122
KP-R-ECD5C00B1267
Research reportStarting point
How reliably do AI assistants answer news questions?
Public-service broadcasters examine errors, source use and distortions in AI assistants' answers.
Why read it?
Provides research against which to assess expectations of AI as a source of news and facts.
Keep in mind
The study concerns particular models, questions and a measurement period. Other systems or situations may produce different results.
Reference details
Title as recorded: Hoe betrouwbaar beantwoorden AI-assistenten nieuwsvragen?
Author(s): Not recorded · Year: Not recorded
Publication / host: www.ebu.ch
Source language: English
KP-R-F782A88F5100
No sources match these filters. Try fewer filters or a different search term.