Sergey Diakov
⬤ Google's Gemini 3 Pro just set a new record in AI trustworthiness. In Prolific's massive HUMAINE evaluation involving 26,000 participants, the model scored an impressive 69% trust rating in blind tests — a huge improvement from Gemini 2.5 Pro's modest 16%. This positions Gemini 3 Pro at the very top when it comes to trust, ethics, and safety.
⬤ The numbers tell a compelling story. Gemini 3 Pro ranked first 69% of the time across different demographic groups, while its predecessor only managed that top spot 16% of the time. The HUMAINE framework tested four key areas, and Gemini 3 crushed three of them: performance and reasoning, interaction and adaptiveness, and trust and safety. It only fell short in communication style, where DeepSeek V3 grabbed 43% of user preferences.
⬤ What's really striking is how consistently Gemini 3 Pro performed across 22 different user groups — spanning age, gender, ethnicity, and political views. When users were shown both models side-by-side without knowing which was which, they picked Gemini 3 five times more often. That kind of consistency across such diverse groups shows real improvement in how people perceive the model's reliability.
⬤ This trust breakthrough represents a significant moment for Google in the AI race. As companies and users increasingly prioritize safety and dependability alongside raw performance, Gemini 3 Pro's results could influence everything from enterprise adoption decisions to market positioning in the rapidly evolving AI landscape.
Sergey Diakov
