Brett Levenson left the tech giant Apple in 2019 with a mission: to improve Facebook's content moderation using better technology. He soon realised that the problem went beyond algorithms; human reviewers were struggling with outdated policies and time constraints.

The process was flawed, often leading to decisions just as accurate as flipping a coin, but days too late for effective action. This issue has only worsened with the rise of AI chatbots, which have produced incidents such as self-harm guidance being provided to teens or harmful imagery bypassing safety filters.

Levenson’s frustration led to the creation of Moonbounce, a company that uses machine learning to enforce policies in real-time. By turning static policy documents into executable logic, Moonbounce ensures content is reviewed and acted upon swiftly and accurately. The technology now serves over 100 million daily active users across various industries, from dating apps to AI companions.

“Safety isn’t just a feature anymore; it’s an integral part of the product,” Levenson explains. “Our customers are now using this technology to make safety a differentiator in their offerings.”