A novel approach to AI model safety is gaining traction, as startup CTGT tackles the pressing issue of unreliable AI outputs. Founded by Cyril Gorlla, who taught himself to code as an immigrant, and Trevor Tuttle, the company is on a mission to empower organizations to deploy AI more thoughtfully.

CTGT's auditing platform identifies biased outputs and hallucinations from models, addressing the root causes of these issues. This is crucial, as a recent poll found that reliability is a top concern for enterprises adopting AI apps. The startup's mathematically-guaranteed interpretability technique differs from existing methods, providing a more performant solution.

Backed by notable investors, including Mark Cuban, CTGT has already secured customers from the Fortune 10 brands. The company's focus on building a robust engineering team and refining its platform could position it for significant growth in the burgeoning market for AI interpretability, projected to reach $16.2 billion by 2028.