Twitter Algorithm Amplifies Anger and Polarization, Study Finds

Thumbnail Image

The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.

A study by Cornell and UC Berkeley researchers found that Twitter's AI-driven recommendation algorithm, especially under Elon Musk's ownership, amplifies tweets expressing anger and political polarization. This has led to increased negative emotions among users and significantly impacted public discourse, contributing to greater social division and animosity.[AI generated]

Why's our monitor labelling this an incident or hazard?

The article explicitly discusses the role of Twitter's algorithm, an AI system that personalizes content feeds, in amplifying hateful and emotionally charged tweets. This amplification has led to increased exposure to hate speech and political animosity, which constitutes harm to communities. Since the AI system's use has directly led to this harm, the event qualifies as an AI Incident under the OECD framework.[AI generated]
AI principles
AccountabilityHuman wellbeingSafetyTransparency & explainabilityDemocracy & human autonomy
Industries
Media, social platforms, and marketing
Affected stakeholders
ConsumersGeneral public
Harm types
PsychologicalPublic interest
Severity
AI incident
Business function:
Marketing and advertisement
AI system task:
Organisation/recommenders

Articles about this incident or hazard

Thumbnail Image

Twitter y la mala leche: un estudio señala que con Elon Musk se han incrementado los tuits de odio

2023-06-01
LaVanguardia
Why's our monitor labelling this an incident or hazard?
The article explicitly discusses the role of Twitter's algorithm, an AI system that personalizes content feeds, in amplifying hateful and emotionally charged tweets. This amplification has led to increased exposure to hate speech and political animosity, which constitutes harm to communities. Since the AI system's use has directly led to this harm, the event qualifies as an AI Incident under the OECD framework.
Thumbnail Image

Estudio revela que el algoritmo de Twitter promueve el contenido de ira

2023-06-03
El Comercio Perú
Why's our monitor labelling this an incident or hazard?
The event involves an AI system explicitly described as Twitter's machine learning algorithm that filters and recommends content. The study demonstrates that this AI system's use directly leads to increased anger and polarization among users, which constitutes harm to communities and public discourse. The harm is realized and significant, not merely potential. Hence, the event meets the criteria for an AI Incident because the AI system's use has indirectly led to social and emotional harm through content amplification and user behavior influence.
Thumbnail Image

¿Ira, odio, amor o felicidad? El algoritmo de Twitter está diseñado para promover uno de estos sentimientos, según estudio

2023-06-02
Semana.com Últimas Noticias de Colombia y el Mundo
Why's our monitor labelling this an incident or hazard?
The Twitter algorithm is an AI system that filters and recommends content based on machine learning. The study shows that it amplifies anger and polarizing political content, which affects users' emotions and behavior, potentially increasing societal division and emotional harm. While no direct harm is reported as having occurred, the amplification of such content plausibly leads to harm to communities and social cohesion. Hence, this is an AI Hazard due to the credible risk of future harm stemming from the AI system's use and effects.
Thumbnail Image

El algoritmo de Twitter amplifica el contenido que expresa ira y polarización

2023-06-03
La Voz de Galicia
Why's our monitor labelling this an incident or hazard?
The event involves an AI system (Twitter's recommendation algorithm) that filters and promotes content based on learned patterns, which is a form of machine learning. The study shows that this AI system's use leads to increased anger and polarization among users, impacting public discourse. This constitutes harm to communities by fostering social division and negative emotional states. Since the harm is occurring as a direct consequence of the AI system's use, this qualifies as an AI Incident.
Thumbnail Image

El algoritmo de Twitter amplifica el contenido que expresa ira y...

2023-06-02
europa press
Why's our monitor labelling this an incident or hazard?
The Twitter algorithm is an AI system that filters and ranks content for users. The study shows that its use leads to increased anger and polarization among users, which constitutes harm to communities (harm category d). Although no physical injury is reported, the social harm from amplified negative emotions and political polarization is significant and directly linked to the AI system's operation. Therefore, this qualifies as an AI Incident due to the realized harm caused by the AI system's use.
Thumbnail Image

El algoritmo de Twitter amplifica el contenido que expresa ira y polarización, según un estudio

2023-06-02
Revista Proceso
Why's our monitor labelling this an incident or hazard?
The study explicitly identifies Twitter's algorithm, which is a machine learning-based AI system, as amplifying harmful emotional content (anger and polarization). This amplification has caused increased negative emotions among users and impacts public discourse, which constitutes harm to communities. The AI system's use is directly linked to these harms, fulfilling the criteria for an AI Incident. The event is not merely a potential risk or a complementary update but documents realized harm resulting from the AI system's operation.
Thumbnail Image

Los cambios en los algoritmos de Twitter de Elon Musk "amplifican la ira y la animosidad", según los investigadores

2023-05-31
es-us.finanzas.yahoo.com
Why's our monitor labelling this an incident or hazard?
The article explicitly discusses the role of Twitter's algorithm, which is an AI system that filters and selects content for users. The study finds that this AI system amplifies emotional and partisan content, causing users to experience more anger and increasing social division. This constitutes harm to communities through increased animosity and polarization, fulfilling the criteria for an AI Incident. The harm is realized and ongoing, not merely potential, as users' emotional responses and behaviors are directly influenced by the AI system's outputs.