Persistent interaction patterns across social media platforms and over time

Nature, Год журнала: 2024, Номер 628(8008), С. 582 - 589

Опубликована: Март 20, 2024

Abstract Growing concern surrounds the impact of social media platforms on public discourse 1–4 and their influence dynamics 5–9 , especially in context toxicity 10–12 . Here, to better understand these phenomena, we use a comparative approach isolate human behavioural patterns across multiple platforms. In particular, analyse conversations different online communities, focusing identifying consistent toxic content. Drawing from an extensive dataset that spans eight over 34 years—from Usenet contemporary media—our findings show conversation user behaviour, irrespective platform, topic or time. Notably, although long consistently exhibit higher toxicity, language does not invariably discourage people participating conversation, necessarily escalate as discussions evolve. Our analysis suggests debates contrasting sentiments among users significantly contribute more intense hostile discussions. Moreover, persistence three decades, despite changes societal norms, underscores pivotal role behaviour shaping discourse.

Язык: Английский

---

Quantifying the impact of misinformation and vaccine-skeptical content on Facebook

Science, Год журнала: 2024, Номер 384(6699)

Опубликована: Май 30, 2024

Low uptake of the COVID-19 vaccine in US has been widely attributed to social media misinformation. To evaluate this claim, we introduce a framework combining lab experiments (total N = 18,725), crowdsourcing, and machine learning estimate causal effect 13,206 vaccine-related URLs on vaccination intentions Facebook users ( ≈ 233 million). We that impact unflagged content nonetheless encouraged skepticism was 46-fold greater than misinformation flagged by fact-checkers. Although reduced predicted significantly more when viewed, users’ exposure limited. In contrast, stories highlighting rare deaths after were among Facebook’s most-viewed stories. Our work emphasizes need scrutinize factually accurate but potentially misleading addition outright falsehoods.

Язык: Английский

---

Misunderstanding the harms of online misinformation

Nature, Год журнала: 2024, Номер 630(8015), С. 45 - 53

Опубликована: Июнь 5, 2024

Язык: Английский

---

The effects of Facebook and Instagram on the 2020 election: A deactivation experiment

Proceedings of the National Academy of Sciences, Год журнала: 2024, Номер 121(21)

Опубликована: Май 13, 2024

We study the effect of Facebook and Instagram access on political beliefs, attitudes, behavior by randomizing a subset 19,857 users 15,585 to deactivate their accounts for 6 wk before 2020 U.S. election. report four key findings. First, both deactivation reduced an index participation (driven mainly online). Second, had no significant knowledge, but secondary analyses suggest that it knowledge general news while possibly also decreasing belief in misinformation circulating online. Third, may have self-reported net votes Trump, though this does not meet our preregistered significance threshold. Finally, effects affective issue polarization, perceived legitimacy election, candidate favorability, voter turnout were all precisely estimated close zero.

Язык: Английский

---

Beyond online disinformation: assessing national information resilience in four European countries

Humanities and Social Sciences Communications, Год журнала: 2024, Номер 11(1)

Опубликована: Янв. 13, 2024

Abstract As social media is a key conduit for the distribution of disinformation, much literature on disinformation in elections has been focused internet and global platforms. Literature societal trust also grown recent years. Yet, not limited to platforms or internet, traditional outlets many European countries act as vehicles often under direction government. Moreover, connection between resilience less discussed. This article aimed at tackling question what makes country vulnerable resilient against online disinformation. It argues that society’s information can be viewed combination structural characteristics, features its knowledge-distribution institutions including system, activities capabilities citizens. The this argument by describing these dimensions four case countries, based comparable statistics document analyses. results indicate European-wide strategies do uniformly strengthen national anti-disinformation need anchored targeted assessments state level more effective. Such are central, particularly understanding citizens’ needs democratic events such elections.

Язык: Английский

---

The Economics of Social Media

Journal of Economic Literature, Год журнала: 2024, Номер 62(4), С. 1422 - 1474

Опубликована: Дек. 1, 2024

We provide a guide to the burgeoning literature on economics of social media. first define media platforms and highlight their unique features. then synthesize main lessons from empirical organize them around three stages life cycle content: (i) production, (ii) distribution, (iii) consumption. Under we discuss how incentives affect content produced off harmful is moderated. network structure, algorithms, targeted advertisements. consumption, affects individuals who consume its society at large, explore consumer substitution patterns across platforms. Throughout guide, examine case studies deterrence misinformation, segregation, political advertisements, effects outcomes. conclude with brief discussion future (JEL D12, D72, D83, D91, I31, L82, M37)

Язык: Английский

---

Embedding Democratic Values into Social Media AIs via Societal Objective Functions

Proceedings of the ACM on Human-Computer Interaction, Год журнала: 2024, Номер 8(CSCW1), С. 1 - 36

Опубликована: Апрель 17, 2024

Mounting evidence indicates that the artificial intelligence (AI) systems rank our social media feeds bear nontrivial responsibility for amplifying partisan animosity: negative thoughts, feelings, and behaviors toward political out-groups. Can we design these AIs to consider democratic values such as mitigating animosity part of their objective functions? We introduce a method translating established, vetted scientific constructs into AI functions, which term societal demonstrate with application science construct anti-democratic attitudes. Traditionally, have lacked observable outcomes use train models-however, sciences developed survey instruments qualitative codebooks constructs, precision facilitates translation detailed prompts large language models. apply this create attitude model estimates extent post promotes attitudes, test across three studies. In Study 1, first attitudinal behavioral effectiveness intervention among US partisans (N=1,380) by manually annotating (alpha=.895) posts scores testing several feed ranking conditions based on scores. Removal (d=.20) downranking (d=.25) reduced participants' without compromising experience engagement. 2, scale up manual labels creating model, finding strong agreement (rho=.75). Finally, in 3, replicate 1 using instead its impact (N=558), again find function (d=.25). This presents novel strategy draw theory methods mitigate harms AIs.

Язык: Английский

---

The Economics of Social Media

SSRN Electronic Journal, Год журнала: 2024, Номер unknown

Опубликована: Янв. 1, 2024

We review the burgeoning literature on economics of social media, which has become ubiquitous in modern economy and fundamentally changed how people interact. first define media platforms isolate features that distinguish them from traditional other digital platforms. then synthesize main lessons empirical organize around three stages life cycle user-generated content: (1) production, (2) distribution, (3) consumption. Under we discuss incentives affect content produced off harmful is moderated. network structure, algorithms, targeted advertisements. consumption, affects individuals who consume its society at large, consumer substitution patterns across Throughout review, delve into case studies examining deterrence misinformation, segregation, political advertisements, effects outcomes. conclude with a brief discussion future media.

Язык: Английский

---

Engagement, user satisfaction, and the amplification of divisive content on social media

PNAS Nexus, Год журнала: 2025, Номер 4(3)

Опубликована: Фев. 27, 2025

Abstract Social media ranking algorithms typically optimize for users’ revealed preferences, i.e. user engagement such as clicks, shares, and likes. Many have hypothesized that by focusing on these may exacerbate human behavioral biases. In a preregistered algorithmic audit, we found that, relative to reverse-chronological baseline, Twitter’s engagement-based algorithm amplifies emotionally charged, out-group hostile content users say makes them feel worse about their political out-group. Furthermore, find do not prefer the tweets selected algorithm, suggesting underperforms in satisfying stated preferences. Finally, explore implications of an alternative approach ranks based preferences reduction angry, partisan, content, but also potential reinforcement proattitudinal content. Overall, our findings suggest greater integration into social could promote better online discourse, though trade-offs warrant further investigation.

Язык: Английский

---

On the impossibility of breaking the echo chamber effect in social media using regulation

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Янв. 11, 2024

Abstract As scientists, we are proud of our role in developing the current digital age that enables billions people to communicate rapidly with others via social media. However, when things go wrong, also responsible for taking an ethical stand and trying solve problems, this work aims take a step direction. Our goal is set foundation mathematically formal study how might regulate media and, particular, address problem echo chamber effect. An closed system where other voices excluded by omission, causing your beliefs become amplified or reinforced. In turn, these bubbles can boost polarization extreme political views, unfortunately, there strong evidence chambers exist The fundamental question try answer is: regulation “break” reduce effect media? Sadly, paper’s main result impossibility result: general function achieves (on model) while obeying core values democratic societies (freedom expression user privacy) does not exist. This leaves us hard future choices make.

Язык: Английский

---