Intelligence Raises 7.9 Million Dollars to Teach AI Models Human Design Preferences

A new startup named Intelligence has secured 7.9 million dollars in seed funding to help artificial intelligence better understand what people truly like in designs and creative outputs.

The company created DesignArena as a platform where users rank AI generated options through simple comparisons to guide model improvements.

The Growing Need for Human Input in AI Training

This funding comes as many AI systems still produce results that are technically correct but lack the appeal that makes them enjoyable or useful in practice.

Co founder Grace Li started the effort with college friends in 2025 while trying to build fun AI powered games that initially fell short on engagement.

The platform now reaches 5.3 million users globally who provide ongoing feedback that helps refine media generating models for companies.

Enterprise clients pay for access to this crowd sourced data because it offers real world insights that automated tests often miss.

Future Outlook for Personalized and Culturally Aware AI Tools

By logging user preferences across different regions the system can track how tastes vary such as more bold styles popular in parts of Asia.

Such regional data adds important context that could help AI avoid generic outputs and better reflect diverse global design trends.

Similar efforts in the space have shown both promise and challenges with some competitors shutting down despite early traction.

Looking ahead this kind of feedback loop may enable AI to create more tailored experiences in fields like gaming web design and visual content.

For regular people improved AI models mean future tools will generate images websites and games that feel more natural and satisfying to use.