Public transcript

Are you creepy? - The psychology of creepiness explained | Julia Shaw and Lex Fridman

By: Lex Clips
Length: 11 min
Language: English
Text from: the video's captions
Transcribed: 12 October 2026

Read the full transcript Open on YouTube

Notes in English.

Summary

Criminal psychologist Julia Shaw talks with Lex Fridman about why our intuitions about creepiness and trustworthiness are unreliable, and how overconfidence in lie detection can lead to wrongful convictions. They discuss psychopathy, tailored treatments, and whether studying 'evil' people carries a psychological cost. Shaw argues for seeing cases as puzzles and for nuance over moral panic.

Key points

Creepiness is largely a reaction to people not following social norms, not an indicator of danger.
People physically and psychologically distance themselves from those with severe mental illness, even though research does not show they are more violent.
Even experienced police officers are no better than chance at detecting lies, despite high confidence in their abilities.
Overconfidence in lie detection can contribute to wrongful convictions and also fuels jealousy in relationships.
Some research suggests people with psychopathy are better at lying, including 'faking good' in parole decisions.
Empathy-focused treatment for psychopathy may be weaponized; Jennifer Skeem's programs instead teach that being pro-social is a better life strategy.
Shaw treats heinous cases as puzzles for pattern recognition rather than emotionally investing in victims, which she sees as necessary for objectivity.
Shaw questions whether covering murderers who kill for fame should be a blanket ban, arguing cases need thoughtful handling rather than silence.

Questions it answers

00:00Creepiness radar

Why do we perceive some people as creepy, and is that perception accurate?

Creepiness is tied to people not following social norms, and our intuitive judgments about it are unreliable. We distance ourselves from people who act erratically, like those with severe mental illness, even though they are not actually more dangerous.

Creepiness judgments are based on deviation from social norms.
People physically distance themselves from those with severe mental illness despite no increased danger.

02:02Attractiveness and deviation

Does attractiveness make someone seem less creepy?

It depends—being too attractive can also make people less trusted. Any deviation from the norm can make our assessments more wrong and more negative.

Attractiveness effects interact in complex ways.
Deviation from the norm in any direction skews judgment.

02:58Lie detection failure

Can trained professionals detect lies better than the rest of us?

No. Even experienced police officers and investigative interviewers, despite high confidence, perform no better than chance in experimental settings. This overconfidence can lead to wrongful convictions.

Aldert Vrij's research shows lie detection is unreliable.
Overconfident lie detectors can wrongly accuse innocent people.

04:19Jealousy and lie detection

How does poor lie detection affect relationships?

Jealousy is often born out of uncertainty—an assumption about what a partner might be thinking or doing. It is essentially an exercise in lie detection, at which we are very bad.

Jealousy stems from uncertainty, not confirmed wrongdoing.

04:39Psychopathy and treatment

Are people with psychopathy better liars, and how should they be treated?

Some research suggests people with psychopathy lie better, including faking rehabilitation in parole decisions. Empathy-focused treatment may be weaponized, so researchers like Jennifer Skeem design programs teaching that pro-social behavior is a better life strategy.

People with psychopathy can 'fake good' in parole hearings.
Tailored treatments focused on pro-social incentives may work better than empathy training.

06:12Studying evil

Is there a psychological cost to studying 'monsters'?

Shaw doesn't fear becoming evil through her work. She sees cases as puzzles for pattern recognition, staying objective while working with police and lawyers rather than emotionally investing in victims.

The 'abyss' is not a blank—recognizable patterns help legal decision-making.
Shaw views cases as puzzles, leaving emotional support work to therapists.

09:05Beauty in the puzzle

What is the risk of finding everyone interesting as a researcher?

Shaw worries that seeing the beauty in every puzzle, like Dostoevsky's Prince Myshkin, could make you lose your footing on the moral landscape. Scientists' delight in discovery can also read as callousness to others.

Fascination with everyone can blur moral judgment.
Scientific curiosity can be misread as callousness.

10:05Nuance and platforming

How should we cover people who commit crimes for fame?

Shaw argues that nuance and sympathy toward written-off people draws criticism but is not trivializing. On true crime coverage of murders committed for fame, she says it should not be a blanket ban, but cases need careful thought about how and whom to cover.

Rejecting the word 'evil' is about understanding, not trivializing.
Media coverage of fame-seeking killers requires careful judgment, not censorship.

Notes

Creepiness and social norms

Julia Shaw explains that creepiness has only recently been defined and is largely tied to people not following social norms.
Our "creepiness radars" are unreliable: we distance ourselves from people with severe mental illness (e.g., sitting more chairs away in a waiting room study), even though research shows they are not inherently more violent or dangerous.
Attractiveness interacts with creepiness judgments in complicated ways; being too attractive can also reduce trust.

Lie detection failures

Research by Aldert Vrij and others shows it is very hard to reliably detect lies.
Police officers and investigative interviewers are highly confident in their lie-detection abilities, but in experimental settings they perform no better than chance.
This overconfidence can lead to missing real liars and to wrongful convictions of innocent people.
Jealousy in relationships is often born of uncertainty and is essentially a failed exercise in lie detection.

Psychopathy and treatment

Some research supports that people with psychopathy are better at lying, including "faking good" in parole decisions.
There is historical concern that empathy-focused treatment for psychopathy can be weaponized—people learn to fake empathy.
Jennifer Skeem's treatment programs focus instead on teaching the rules of society and convincing people that being pro-social is a better way to get what they want.

Studying evil

Shaw rejects the idea that studying "monsters" makes you one; the "abyss" is not a blank—there are patterns to recognize that help legal decision-makers.
She treats cases as puzzles and does pattern recognition rather than emotionally investing in victims; therapists do that work.
She worries about losing moral footing by seeing "the beauty in the puzzle" in everyone, comparing herself to Prince Myshkin from Dostoevsky's The Idiot.
Scientists' delight in discovery can read as callousness, and nuance can estrange others—e.g., her argument against using the word "evil."
On true crime media, she raises questions about covering murders committed for fame: not a blanket ban, but careful thought about how and whom to cover.

Text from the video's own captions. Summary and notes written by AI from the transcript, so check anything important against the recording.