FrameCheck
A browser extension that reads articles, social posts, and AI-generated content alongside you and tags each passage's epistemic stance—reported fact, opinion, satire, AI-generated, or fictional—so you can calibrate trust as you scroll.
News consumers and parents worried about AI-generated content
- Inline epistemic labels (Fact, Opinion, Satire, AI-generated, Fiction) with hover explanations
- Trust-meter summary for the page highlighting framing inconsistencies
- Reading mode that converts dense news into 'claim → evidence → confidence' view
- Shareable receipts when forwarding articles so friends see the framing notes
Frontier LLMs demonstrably fail to flag fictional or misaligned content they were trained on (the source paper shows only 9% recognition), and consumers are now drowning in a mix of human and AI-generated writing with no honest labels.
At least 5 active Chrome/Edge extensions already market AI/misinformation detection (TruthCheck.AI, Truthly, FactChecker Pro, Facticity, AI Fact-Check Extension), and marketing copy like FactChecker Pro's 'In 2026, AI-generated content makes it nearly impossible to tell real from manufactured' confirms strong end-user pull.TruthCheck.AI - AI-Powered Fact Checking Chrome Extension ↗FactChecker Pro - Microsoft Edge Add-ons ↗
Direct per-passage epistemic tagging is differentiating, but the 'badge what you read' browser-extension niche already has crowded incumbents (Ground News bias extension, AllSides Bias Finder, Truthly, TruthCheck) serving the same scrolling user; FrameCheck's stance taxonomy is a refinement, not a wide-open lane.Want a second opinion on the news that you're reading? - Ground News ↗Apps & Extensions | AllSides ↗Truthly - Authenticity Detector ↗
Ground News proves willingness to pay with $29.99/year Premium subscriptions plus tiered upgrades; comparable AI fact-check extensions are launched with paid plans, and parents/news consumers are a documented paying segment for media literacy tools.Ground News Subscription ↗How Much Does Ground News Cost? - The Pricer ↗
AI-generated content volume is structurally rising, regulatory pressure for content labels (EU AI Act watermarking) is expanding, and the underlying 'negation neglect' problem from the cited arXiv paper shows model misalignment on epistemic frames is persistent, not a fad.[2607.01690] Epistemic Goggles: A Pretrained Module that Induces an ... ↗FactChecker Pro - Microsoft Edge Add-ons ↗
Browser extension shell is straightforward, but real-time per-passage classification across five epistemic categories (fact/opinion/satire/AI-generated/fictional) requires an LLM call or fine-tuned model per article, raising latency and API cost; satire-vs-opinion and AI-vs-human are notoriously low-accuracy tasks, so engineering effort is non-trivial.AI Fact-Check Extension - Chrome Web Store ↗