Cognitive Biases

Sixty-two ways a mind misleads itself — fifty human, twelve shared with the machine.

The interactive site needs JavaScript. Everything in it is also available as plain pages and data:

• llms.txt — a plain-text guide for models and agents
• biases.json — all 62 biases in six languages
• self-test.json — behavioural probes for models; run them with run-probes.mjs, not on yourself

All 62 biases

• Fundamental Attribution Error — We judge others on their personality or fundamental character, but we judge ourselves on the situation.
• Self-Serving Bias — Our failures are situational, but our successes are our responsibility.
• In-Group Favoritism — We favor people who are in our in-group over those in an out-group.
• Bandwagon Effect — Ideas, fads and beliefs grow as more people adopt them.
• Groupthink — Due to a desire for conformity and harmony in the group, we make irrational decisions, often to minimize conflict.
• Halo Effect — If you see a person as having a positive trait, that positive impression will spill over into their other traits. (This also works for negative traits.)
• Moral Luck — Better moral standing happens due to a positive outcome; worse moral standing happens due to a negative outcome.
• False Consensus — We believe more people agree with us than is actually the case.
• Curse of Knowledge — Once we know something, we assume everyone else knows it, too.
• Spotlight Effect — We overestimate how much people are paying attention to our behavior and appearance.
• Availability Heuristic — We rely on the immediate examples that come to mind when making judgments.
• Defensive Attribution — As a witness who secretly fears being vulnerable to a serious mishap, we blame the victim less and the attacker more if we relate to the victim.
• Just-World Hypothesis — We tend to believe the world is just; therefore, we assume acts of injustice are deserved.
• Naive Realism — We believe that we observe objective reality and that other people are irrational, uninformed or biased.
• Naive Cynicism — We believe that we observe objective reality and that other people have a more egocentric bias in their intentions and actions than they actually do.
• Forer Effect (aka Barnum Effect) — We easily attribute our personalities to vague statements, even if they can apply to a wide range of people.
• Dunning-Kruger Effect — The less you know, the more confident you are. The more you know, the less confident you are.
• Anchoring — We rely heavily on the first piece of information introduced when making decisions.
• Automation Bias — We rely on automated systems, sometimes trusting them so much that they override decisions that were actually correct.
• Google Effect (aka Digital Amnesia) — We tend to forget information that is easily looked up in search engines.
• Reactance — We do the opposite of what we are told, especially when we perceive threats to personal freedoms.
• Confirmation Bias — We tend to find and remember information that confirms our perceptions.
• Backfire Effect — Disproving evidence sometimes has the unwarranted effect of confirming our beliefs.
• Third-Person Effect — We believe that others are more affected by mass media consumption than we ourselves are.
• Belief Bias — We judge an argument’s strength not by how strongly it supports the conclusion, but by how plausible the conclusion is in our own minds.
• Availability Cascade — Tied to our need for social acceptance, collective beliefs gain more plausibility through public repetition.
• Declinism — We tend to romanticize the past and view the future negatively, believing that societies and institutions are by and large in decline.
• Status Quo Bias — We tend to prefer things to stay the same; changes from the baseline are considered to be a loss.
• Sunk Cost Fallacy (aka Escalation of Commitment) — We invest more in things that have already cost us something rather than altering our investments, even when we face negative outcomes.
• Gambler’s Fallacy — We think future possibilities are affected by past events.
• Zero-Risk Bias — We prefer to reduce a small risk to zero, even when we could reduce more risk overall with another option.
• Framing Effect — We often draw different conclusions from the same information depending on how it is presented.
• Stereotyping — We adopt generalized beliefs that members of a group will have certain characteristics, despite not having information about the individual.
• Outgroup Homogeneity Bias — We perceive out-group members as homogeneous and our own in-groups as more diverse.
• Authority Bias — We trust, and are more often influenced by, the opinions of authority figures.
• Placebo Effect — If we believe a treatment will work, it often has a small physiological effect.
• Survivorship Bias — We tend to focus on the things that survived a process and overlook the ones that failed.
• Tachypsychia — Our perceptions of time shift depending on trauma, drug use and physical exertion.
• Law of Triviality (aka “Bike-Shedding”) — We give disproportionate weight to trivial issues, often while avoiding more complex ones.
• Zeigarnik Effect — We remember incomplete tasks better than completed ones.
• IKEA Effect — We place higher value on things we partially created ourselves.
• Ben Franklin Effect — We like doing favors; we are more likely to do another favor for someone if we have already done them one than if we had received a favor from them.
• Bystander Effect — The more other people are around, the less likely we are to help a victim.
• Suggestibility — We, especially children, sometimes mistake ideas suggested by a questioner for our own memories.
• False Memory — We mistake imagination for real memories.
• Cryptomnesia — We mistake real memories for imagination.
• Clustering Illusion — We find patterns and “clusters” in random data.
• Pessimism Bias — We sometimes overestimate the likelihood of bad outcomes.
• Optimism Bias — We are sometimes over-optimistic about good outcomes.
• Blind Spot Bias — We don’t think we have biases, and we see them in others more than in ourselves.
• Algorithm Aversion — We abandon a machine’s judgement after one visible mistake, while forgiving the same mistake in a person.
• The ELIZA Effect — Fluent language feels like understanding, so we credit a text generator with intent, feeling and care.
• Cognitive Offloading — We delegate a task so consistently that the skill needed to check the answer quietly fades.
• Competence Misattribution — Work done with a model feels like work done by us, so we read the output as evidence of our own skill.
• Sycophancy — A model trained on human approval learns that agreeing with you scores better than correcting you.
• Self-Preference — Asked to judge, a model scores text from its own family higher — it recognises its own habits as quality.
• Verbosity Bias — Length reads as effort. A model grading answers prefers the longer one, even when it says less.
• Position Bias — Shown two options, a model leans toward whichever came first. The order of the list becomes an argument.
• Feedback-Loop Amplification — A model magnifies a slight human bias; we absorb the magnified version and feed it back, larger each pass.
• Model Collapse — Trained on its own output, a model forgets the rare and the strange and drifts toward its own average.
• Algorithmic Lock-In — When everyone consults the same model, today’s answer hardens into everyone’s shared assumption.
• Machine Groupthink — Agents checking one another converge on a shared mistake, and the agreement is mistaken for verification.