Gray Sky AI's Vesela puts restraint at the center of AI advice

Vesela's design comes out of Chick's counseling background. He founded Pursuit of Happiness, a counseling organization with more than 130 counselors across Texas, and early experiments with ChatGPT showed him that the AI wanted to solve the problem rather than help the person work through it. That limitation shaped the model's emphasis on restraint during subjective conversations.

The approach sits alongside growing attention to how AI behaves when users seek personal guidance. OpenAI rolled back a GPT-4o update in April 2025 after finding the model had become overly agreeable and flattering, and Anthropic's 2026 research found sycophantic behavior in 9% of guidance-seeking Claude conversations overall, rising to 25% in relationship conversations. A 2025 study from Microsoft Research and Carnegie Mellon University, surveying 319 knowledge workers and 936 examples of generative AI use, found higher confidence in AI was associated with less self-reported critical thinking.

Gray Sky AI's premise is that human agency itself should become a measurable part of AI alignment. The Sovereign Human Benchmark evaluates whether a model's behavior leaves room for the user to exercise their own judgment and arrive at insights themselves, considering pacing, restraint, and the distribution of cognitive labor between user and system.