Observed Signal · Jun 6, 2026 · Technical Release · Source: DEV Community · Impact: 2/5 · Sentiment: Positive
Prototype Chrome Extension ClassifierAI Detects AI-Generated Content
Two developers, Francis Tran and Elmar Chavez, published ClassifierAI: an open-source prototype Chrome extension that uses TensorFlow.js to classify whether dev.to articles and their cover images are AI-generated. The extension performs image classification (trained via Google Teachable Machine on a Hugging Face dataset of 866 images: 433 AI / 433 NotAI) and text classification (using a Wikipedia-based human/AI dataset). It runs in the browser (Manifest V3), provides a UI with icons and a popup showing classification results, and includes installation/build instructions and the GitHub repository. The authors note accuracy limitations because datasets are not tailored to dev.to and list current challenges (dataset mismatch, limited per-section attribution). The project was built during a GitHub Finish-Up-A-Thon challenge and is presented as a community-tuneable prototype.
An open-source, in-browser ML prototype addresses detection of AI-generated content which is relevant to content quality and moderation, but its impact is limited by dataset mismatch and prototype status.
Track GitHub Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Francis Tran (Main Author) and Elmar Chavez (Co-Author) built ClassifierAI, an open-source prototype Chrome extension.
- ClassifierAI uses TensorFlow.js for both image classification and text classification in-browser (Chrome Extension Manifest V3).
- Image classifier trained with Teachable Machine on a Hugging Face dataset of 866 images (433 AI, 433 NotAI); training settings: 30 epochs, batch size 16, learning rate 0.0001.
- Text classification uses the Hugging Face dataset gouwsxander/wikipedia-human-ai and scores articles as Human-written, AI-generated, or Mixed.
- Repository and installation instructions provided: git clone https://github.com/FrancisTR/ClassifierAI.git and build/run as an unpacked Chrome extension.
Connected Companies & Entities
2 Entities mappedOntology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Pangram launches Chrome extension to label AI content
Pangram, a startup, launched a Chrome extension that labels AI-generated or AI-assisted text in real time across social and publishing feeds (X, LinkedIn, Substack, Reddit, Medium). The extension provides inline labels, a sidebar tracking the share of AI content in a user’s feed, and a right-click scan for selected text. Pangram says its token-level classifier — trained on millions of paired human/AI documents and calibrated to strongly penalize false positives — yields a claimed 1-in-10,000 false positive rate and enforces a 50-word minimum for reliable detection. The company collaborated with the Internet Archive on prevalence estimates and with NewsGuard on identifying “AI Content Farm” sites. Pangram’s commercial bet is a $20/month consumer subscription, with expansion plans (image/video detection and potential API use) discussed as possible growth paths.
Chrome Extension 'Slop Mop' Cleans LinkedIn Feed of AI-Generated Posts
A developer has created 'Slop Mop', a Chrome browser extension designed to detect and hide AI-generated content in LinkedIn feeds. The tool uses a decision model called Jev from Typesafe to analyze posts and identifies nine signs of AI slop, including hype words like 'game-changing' and 'unlock', as well as generic phrasing patterns. If a post reaches a 70% AI probability threshold, it can be hidden or flagged with a red mop icon. Users can also manually rate posts to improve the system's accuracy. This comes after LinkedIn introduced a reporting tool for AI-generated content, which has flagged over a million posts. The extension has been tested but shows mixed accuracy, such as marking a post explicitly stating it was created with ChatGPT as only 54% AI-probable. The tool is not an official LinkedIn feature but a community-driven solution.
UltraHumanizer.ai Launches Free AI Image Detector
UltraHumanizer.ai, an AI-powered content refinement platform, has launched a completely free AI Image Detector. The tool allows users to upload an image and receive a probability-based likelihood that generative AI was involved in creating or editing it. Unlike binary classifiers, it presents results as estimates, reflecting limitations such as partial AI editing, compression, and newer AI models. This expansion moves the platform beyond its original AI humanizer for text, now covering both text refinement and image analysis. The detector is available without a subscription, aligning with the company's free-access approach. It targets content teams, marketers, publishers, and creators looking to add an extra layer of information when reviewing digital images.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.