M Data Launches Video Metadata Generation Service Combining Generative AI with Human Expertise

M Data, a company specializing in the research, analysis, and distribution of TV program and commercial data, has launched a service designed to generate and enrich metadata for video content using generative AI, complemented by its own data correction skills and metadata generation expertise accumulated over 20 years. By utilizing its 'TV Metadata Library' and expert contextual interpretation, M Data aims to solve practical issues that operators face when adopting AI.

While many streaming providers are using generative AI and image analysis to automate metadata generation, they face significant challenges such as AI hallucinations, misrecognition of proper nouns, and exorbitant processing costs. M Data addresses these issues through a 'Human-in-the-Loop' approach, incorporating human expertise into AI-driven workflows.

Key solution points:
- Accurate generation by inputting existing high-quality metadata as contextual information for AI.
- Hybrid processing where expert operators check and supplement AI-generated data.
- Cost reduction by pinpointing specific exposure locations, thereby limiting AI processing to necessary segments.

Three core values provided:
- DX Acceleration: Evolving metadata input into AI-mainstream workflows to increase efficiency.
- Problem Solving: Correcting AI weaknesses regarding jargon, proper nouns, and context interpretation.
- Future Value Creation: Monetizing video assets through scene segmentation, summaries, and advertising metadata.

The service is designed to integrate flexibly with existing AI and data platforms, and is being promoted as 'TV Metadata MCP'.