Why did Nicholas Christakis stop using Twitter?

Nicholas Christakis found Twitter to be incredibly toxic, filled with garbage, trolling, and far-right conspiracy theories. The platform's algorithms also led to a feed that became primarily 'AI slop' rather than useful information.

What is 'AI slop' and why is it a problem?

AI slop refers to low-quality, often fictional or misleading content generated by AI that floods online platforms. It's a problem because it degrades the information environment, making it difficult to discern reality from fiction and potentially eroding trust in information sources.

How can social media harms be addressed?

Potential solutions include lawsuits against social media companies, revocation of Section 230, a shift in social norms towards more responsible usage, and potentially the inherent limitations of AI-generated content making people value reputable sources more.

What is the dual nature of AI's development?

Experts debate AI's future, with some highlighting its extraordinary promise for a better future, while others warn of significant risks, including a percentage chance of human extinction.

How can 'dumb' AI be used to improve human interaction?

The research focuses on using AI not to replace human cognition but to act as a catalyst, like platinum in a chemical reaction. Thoughtful injection of AI agents can optimize human collective and individual performance by facilitating interactions.

What are the potential consequences of interacting with humanoid robots?

Interacting with highly realistic humanoid robots could lead to contamination of both self-perception and how others view individuals, especially if these interactions involve immoral behaviors. Conversely, it might also reintroduce social graces, as seen with polite interactions with LLMs.

Key Moments

What Is Technology Doing to Us?

Sam Harris

Science & Technology5 min read21 min video

Mar 24, 2026|22,187 views|583|98

sam harris sam harris podcast waking up podcast waking up waking up with sam harris sam harris waking up waking up sam harris sam harris jordan peterson sam harris joe rogan author neuroscientist philosopher

Save to Pod

Key Moments

TL;DR

Social media has fueled polarization and mental health crises, while AI slop makes it impossible to discern truth, necessitating a return to reputable sources.

Key Insights

Communication technologies from the last decade have significantly contributed to societal polarization, anomie, and mental health crises.

Anonymity and pseudonymity on online platforms inherently lead individuals to behave worse, a phenomenon observed in mobs and online trolling.

The proliferation of AI-generated 'slop' creates a significant problem where discerning the authenticity of content becomes a constant challenge.

Nicholas Christakis abandoned Twitter due to its toxicity and influx of 'garbage,' including far-right conspiracy theories and 'left craziness.'

The potential for AI to generate content from private user data raises concerns about privacy and the training of algorithms without explicit consent.

There's a potential, and perhaps ironic, future where AI slop accelerates a return to privileging reputable sources, as people may be willing to pay for reliability.

The detrimental impact of recent communication technologies

Nicholas Christakis argues that the communication technologies developed over the past decade have been largely harmful to society. He contends that these technologies have exacerbated societal polarization, contributed to feelings of anomie (a state of normlessness), and worsened mental health crises. Furthermore, these platforms have facilitated the growth of a surveillance state, with technologies being used in ways that verge on totalitarianism. Christakis draws a parallel to past environmental cleanup efforts, suggesting that society may eventually overcome these technological challenges, but it will require a significant period of adjustment, perhaps half a generation. This period is characterized by a regression where society has been "yielded to" and "adversely affected by" these technologies before ultimately overcoming them.

Personal disengagement from toxic platforms like Twitter

Christakis shares his personal experience with social media, having become "very disgusted" with Twitter. Initially, he used it as a valuable source of information and a way to access experts. However, over the last few years, the platform became "incredibly toxic," filled with "garbage," trolling, and a significant amount of far-right conspiracy theories and "left craziness." He found the content unusable and eventually stopped using Twitter, migrating to Blue Sky to access scientific content and have more reasonable interactions. He notes that his follower count on Blue Sky is a tenth of what it was on Twitter, but this is acceptable given the improved quality of engagement.

The pervasive problem of AI-generated 'slop'

A significant concern discussed is the rise of "AI slop" – AI-generated content that is often fictional or misleading. Christakis recounts how algorithms, initially feeding him genuine content like BBC photos of baby elephants, began to generate fabricated scenarios, such as a crocodile attacking a baby elephant. He admits to being initially taken in by this "all fiction." This "slop" is deemed useless by Christakis and poses a serious problem, eroding the trustworthiness of online information. The constant inundation of such content raises fundamental questions about the reality of what is being presented online, making it difficult to distinguish truth from fabrication. The implication is that this content floods and degrades the information ecosystem.

Anonymity and its role in online behavior

The conversation touches upon the well-understood link between anonymity and worse behavior. Historically, people in mobs or those wearing masks have exhibited disinhibited actions. Christakis notes that individuals often behave worse when anonymous or pseudonymous online. While he acknowledges that affording people the opportunity to be non-anonymous could improve behavior on social media, he is hesitant to abolish anonymity entirely, seeing it as a tool against totalitarianism. He suggests that social media companies that allow for and privilege non-anonymous accounts might offer a path forward, potentially a reintroduction of something akin to Twitter's old blue checkmark system.

The debate around Section 230 and platform responsibility

The current legal landscape, including a lawsuit against social media companies in California and discussions around Section 230, is briefly mentioned. Christakis struggles with the concept of Section 230, recognizing its historical importance for the internet's emergence and the argument that social media companies are merely carriers. However, he also acknowledges the problematic nature of platforms washing their hands of content entirely, which enables abuses. He does not possess a definitive answer to this complex issue but suggests that the eventual outcome might involve people becoming more discerning, potentially willing to pay for reliability, and that AI itself might paradoxically accelerate this shift by highlighting the prevalence of unreliable content.

The dual nature of AI's promise and peril

Regarding the broader implications of AI, Christakis expresses a sentiment of "dualism," likening it to the character from Fiddler on the Roof who agrees with opposing viewpoints. He notes the existence of expert computer scientists and tech billionaires who espouse the extraordinary promise of AI, while others, equally credible, warn of existential risks, such as a significant human extinction risk. Sam Harris recalls figures like Sam Altman suggesting a 2% to 20% extinction risk from AI, which Christakis finds "psychotic." This dichotomy in expert opinion makes it challenging to form a firm conclusion about AI's ultimate trajectory.

AI's subtle influence on human interaction and social graces

Christakis posits a "toy model" to illustrate how AI can subtly alter human behavior, using Amazon's Alexa as an example. The design of digital assistants necessitates a direct, non-polite interaction style (e.g., "Alexa, weather"), contrasting with human-to-human politeness. He hypothesizes that children interacting with these machines might learn to be rude, which could then extend to their interactions with other humans. His lab is researching how "dumb AI" (AI designed to supplement, not replace, human cognition) can act as a catalyst to optimize human interactions. Early experiments suggest that thoughtful introduction of AI agents can indeed improve collective and individual human performance. This research implies that even non-sentient AI can shape social dynamics and manners, potentially backfiring into less socially appropriate behavior if not carefully designed.

The potential return to valuing reputable sources

Christakis suggests that the sheer volume of unverified and AI-generated content online might paradoxically lead to a "return to privileging reputable sources." Just as people used to trust established news anchors like Dan Rather, there may be a shift back towards valuing publications like The Economist, where information is perceived as more reliable. This reprivatization of trust could mean individuals are less willing to believe whatever they see online and are more inclined to seek out and potentially pay for content from trusted, established authorities. This trend could be accelerated by the very "AI slop" that currently makes discerning truth so difficult, by highlighting the need for verifiable information.

Mentioned in This Episode

●Products

●Software & Apps

●People Referenced

Common Questions

Information technology in the last decade has contributed to societal polarization, anomie, mental health crises, and the development of a surveillance state. These technologies have exploited fundamental human desires, leading to adverse effects on society.

Topics

Ai-Ethics Social Media Impact AI & Machine Learning Technology & Innovation Society & Philosophy Future Of AI Information Integrity Online Behavior Human-machine Interaction Technological Self-harm

Mentioned in this video

Companies

Twitter

A social media platform discussed for its toxicity, trolling, and conspiracy theories, leading the speaker to abandon it.

YouTube

A platform where Nicholas Christakis has started a channel to advance public understanding of science.

People

Nicholas Christakis

Director of the Human Nature Lab at Yale, an MD and sociologist who studies the interaction between humans and technology.

Sam Altman

Mentioned as a tech billionaire who discussed the potential human extinction risk from AI, with estimates ranging from 2% to potentially higher.

Paul Bloom

A psychologist with whom Nicholas Christakis has discussed the implications of Westworld and humanoid robots.

Legislation & Policy

Section 230

Legislation related to social media companies' responsibility for user-generated content, debated in the context of potential reforms.

Software & Apps

Alexa

A digital assistant used as an example to illustrate how human-machine interaction can influence human behavior and politeness, especially in children.

LLM

Large Language Models, where the speaker notes a tendency to be inappropriately polite when typing instructions, reflecting a return of social graces.

Media

Fiddler on the Roof

A movie referenced for a scene illustrating conflicting viewpoints, used as a metaphor for the polarized expert debates on AI.

Westworld

A TV series discussed in the context of humanoid robots and the philosophical implications of interacting with indistinguishable artificial beings, particularly regarding psychopathy and moral contamination.