Higher prices during a storm can be a normal response to increased demand, even when consumers perceive the increase as unfair. The concern becomes more serious when sellers capitalize on short-term emotional pressure that makes people willing to pay far more than a product is worth.

Emergency demand can justify a market price adjustment.
People often view crisis-related increases as morally objectionable.
Scarcity pricing and exploitation are not identical.
Emotional intensity can create a behavioral vulnerability.

“though it might be just a sensible market adjustment.”
Cass Sunstein · 16:26

“then there's a behavioral bias which is motivating people's willingness to pay a lot more than the product is worth.”
Cass Sunstein · 17:07

#surge pricing#price gouging#behavioral economics#markets

Hot Take· 2

Hot Take07:56

Personalized Feeds Can Fracture a Democracy's Shared Reality

Algorithmic feeds and large language models can produce sharply different realities for different communities. Sunstein describes this as a significant democratic threat because mutual understanding and collaborative problem-solving depend on some shared informational ground.

Geographic communities may receive radically different accounts of reality.
Algorithms and language models can both drive fragmentation.
Mutual understanding suffers when citizens lack shared facts.
Collective problem-solving becomes harder.

“Really significant?”
Cass Sunstein · 07:56

“And that can be a real problem for understanding one another and also for mutual problem solving.”
Cass Sunstein · 08:16

#democracy#polarization#llms#shared reality

Hot Take17:42

Europe and America May Both Be Regulating the Wrong Problem

Sunstein says the United States is less privacy-focused than Europe but questions whether privacy is the central issue in algorithmic markets. He argues that neither jurisdiction has fully targeted systems that exploit information deficits or behavioral biases to induce damaging purchases.

The United States is less privacy-focused than Europe.
More privacy regulation is not necessarily the complete answer.
The central harm can be exploitation rather than data collection itself.
Algorithmic abuse overlaps with fraud and deception.

“I think neither Europe nor the US has put their regulatory finger on the actual problem.”
Cass Sunstein · 17:54

“It's in the same universe as fraud and deception”
Cass Sunstein · 18:16

#privacy#regulation#europe#consumer protection

Explainer· 3

Explainer04:14

How Recommendation Algorithms Can Calcify Cultural Tastes

Recommendation systems can serve people more of what they already enjoy, but this convenience can narrow future discovery. Sunstein argues that repeated similarity-based recommendations may harden tastes and culturally separate audiences into isolated groups.

Preference matching can initially feel liberating and useful.
Similarity-based recommendations reduce exposure to unfamiliar material.
Repeated reinforcement can calcify individual tastes.
Cultural groups can become increasingly separated.

“Now, that might seem not like algorithmic harm, that might seem like a triumph of freedom and markets, but it might mean that piece of…”
Cass Sunstein · 04:38

“And that's culturally damaging, and it's also damaging for the development of individual tastes and preferences.”
Cass Sunstein · 05:06

#recommendations#culture#music#personalization

Explainer06:07

Algorithms Can Build Echo Chambers for Us

Echo chambers no longer arise only from people's deliberate media choices. Engagement-optimizing algorithms can infer a political interest and repeatedly funnel aligned material because clicks generate revenue, separating citizens into incompatible informational worlds.

Earlier echo-chamber concerns focused on individual choices.
Algorithms can automatically intensify ideological filtering.
Engagement incentives reward repeated delivery of aligned content.
Separate information environments undermine learning and self-government.

“What you're pointing to is also emphasized in the book, which is that algorithms can echo chamber you.”
Cass Sunstein · 06:29

“because clicks are money”
Cass Sunstein · 06:44

#echo chambers#news#democracy#engagement

Explainer14:33

Generative AI Scales Both the Beauty and Ugliness of Personalization

Generative AI can move beyond familiar recommendation algorithms by creating individualized interactions in real time. That capacity can help users obtain what they want, but it can also produce highly tailored persuasion designed to exploit a person's susceptibilities.

Generative AI extends beyond conventional recommendation systems.
Personal knowledge can improve assistance and relevance.
The same knowledge enables personalized exploitation.
Investment promotion is especially vulnerable to persuasive targeting.

“here's what you like, here's what you want, we're going to help you”
Cass Sunstein · 14:48

“here's how we can exploit you to get you to buy.”
Cass Sunstein · 14:57

#generative ai#personalization#manipulation#investing

Story· 1

Story13:48

ChatGPT Inferred Scarily Precise Details From Dozens of Prompts

Sunstein recounts asking ChatGPT to describe him after only dozens of interactions and finding some conclusions uncomfortably precise. He warns that language models can learn extensively from prompt histories and potentially connect those inferences with identifiable online information.

Prompt histories can reveal substantial personal information.
Accurate inferences may emerge from only dozens of interactions.
Names can connect inferred traits to wider online data.
Effective privacy protections are needed around model interactions.

“it said a few things that were kind of scarily precise about me”
Cass Sunstein · 14:01

“large language models that track your prompts can know a lot about you”
Cass Sunstein · 14:13

#privacy#chatgpt#profiling#prompt data

Frameworks in this episode

Algorithmic Harm Test

Two-Line Algorithmic Consumer Defense

Insights & moments

⚡Myth Buster· 1

A Price Spike Is Not Automatically Exploitative Gouging

◆Hot Take· 2