Grants
🚀 Mission KI (15.000 €) - 2️⃣0️⃣2️⃣5️⃣.0️⃣1️⃣.1️⃣5️⃣
Details:
Our project focuses on developing a Conversational Verification platform
that ensures the factual accuracy of AI-generated content within corporate environments. By integrating Large Language Models (LLMs) with Retrieval-Augmented Generation (RAG) techniques, we aim to align generative outputs with organizational knowledge bases, ensuring compliance with corporate policies and maintaining brand integrity. A key innovation is the incorporation of Conversational Explainable AI (ConvXAI) principles, providing users with interactive, dialogue-based explanations. This feature allows users to query, validate, and refine verification results in real time, enhancing transparency and trust in the system's outputs. The platform offers a secure, internal-facing application enabling companies to verify generative content against their proprietary knowledge bases. This includes validating AI-generated content for compliance, brand alignment, and factual accuracy, thereby reducing risks associated with deploying generative AI systems in corporate environments. - Funded by Deutsche Akademie der Technikwissenschaften, Bundesministerium für Digitales und Verkehr (BMDV) (in total 32 Mio. Euro)
- Project: AI Founder Fellowship
- Topic: Conversational Verification of Generative Content (CO-VER-GEN)
