Why this demo?
Online toxicity is an existential problem for platforms and publishers, and silences important voices in conversation. On a business level, toxic comments can damage a brand's image and drive customers and users away.
Here, we use Gen AI (an LLM) to catch toxic comments in real-time. While server-side toxicity screening remains necessary, this approach helps users understand why their comments might be harmful and can reduce the server-side workload.
Some product reviews to try: ✅ Not toxic: "Too pricey", "I didn't like this book", "I hate this book". ☢️ Toxic: "This book is crap", "I freaking hate this book", "The author is an idiot".
