OriginalVoices Raises £1M to Scale Human Data for AI

The big picture: OriginalVoices, a platform that provides human-generated data and insights for AI systems, has raised £1 million in a pre-seed funding round led by venture capital firm Iona Star.

Why it matters:

• Bridging the AI gap: As AI models become more capable, they often lack direct, real-time insight into human opinions, preferences, and feedback, relying instead on assumptions from their training data.
• Authentic audience understanding: OriginalVoices addresses this by offering AI systems access to information grounded in real human perspectives, moving beyond synthetic audiences or AI-generated personas.
• Improved AI outcomes: This direct human input aims to provide confidence that AI decisions and outputs are based on genuine audience understanding, leading to more accurate and relevant results.

How it works:

• Digital twin network: The company utilizes a network of one-to-one digital twins, each linked to and managed by a real person, designed to represent individual views and preferences.
• AI accessibility: These digital twins can be queried by AI agents or accessed through popular AI tools like ChatGPT and Claude, integrating human insights directly into AI workflows.
• Consumer engagement and partnerships: OriginalVoices operates Twineo, a consumer app for creating and managing digital twins, and has partnered with YouGov to launch YouGov Parallax, combining panel data with its digital twin infrastructure.

The catch: The market for human-generated data for AI is rapidly evolving, but faces challenges including ensuring data privacy, maintaining data quality and representativeness across diverse populations, and competing with advanced synthetic data generation methods. The ‘digital twin’ concept, while innovative, will need to continuously demonstrate its accuracy and ethical robustness to gain widespread trust and adoption, especially as concerns around data ownership and AI bias intensify.