3 ms·

Show HN: We built a Chrome extension to check facts and detect media bias

2 points by swoopuser 2y ago

Hey HN, We’ve been working on a Chrome extension called Sidecar, designed to make browsing safer and more informed by providing fact-checking, bias detection, and toxicity analysis for AI chats and web content (on any website). The idea: With the rise of generative AI and vast online content, misinformation, bias, and harmful content seem to be everywhere. Sidecar acts as a companion to verify information in real-time. When you highlight text on a webpage, the Sidecar pigeon icon appears, giving you quick access to fact-checks, bias detection, and context. Here’s what Sidecar can do: Instant Verification: Highlight content or click Sidecar to check for inaccuracies, receiving a sourced, accurate breakdown. Bias Detection: Understand content through ratings for bias (political, age, gender) and toxicity. Interactive Corrections: Rewrite or fix inaccurate text based on Sidecar’s analysis. Compatibility: Works with major AI models like ChatGPT. We’d love for you to try it out! You can download Sidecar from the Chrome Web Store (https://chromewebstore.google.com/detail/sidecar-preview-by-autoal/fmdfgdonhkjpdigljddfpgeokdmnehpf https://chromewebstore.google.com/detail/sidecar-preview-by-...). Thanks for checking it out! Feel free to check out our website: https://www.autoalign.ai/sidecar https://www.autoalign.ai/sidecar

codingdave 2y ago
Using AI for fact checking? Sounds dubious. LLMs remain notoriously bad at accurate facts to begin with.
- swoopuser 2y ago
  It's an excellent point and agree its important to fact-check LLM produced content! We're not using LLMs for precisely this reason - we've built this from the ground up to detect issues in LLMs. We're putting together a publication that shows more about the technical approach and also show its performance against both industry and our own evals to highlight the effectiveness. We can share here when its published.

2 points by swoopuser 2y ago

codingdave 2y ago

Using AI for fact checking? Sounds dubious. LLMs remain notoriously bad at accurate facts to begin with.

swoopuser 2y ago
It's an excellent point and agree its important to fact-check LLM produced content! We're not using LLMs for precisely this reason - we've built this from the ground up to detect issues in LLMs. We're putting together a publication that shows more about the technical approach and also show its performance against both industry and our own evals to highlight the effectiveness. We can share here when its published.