Brett Levenson, an entrepreneur and former Apple executive, is creating big buzz in the tech sector through his fresh and creative company Moonbounce. After leaving Apple in 2019, Levenson became the leader of Facebook, where he headed up business integrity. His anger at moderation as a blunt instrument inspired his mission. In response, he launched Moonbounce, a startup responsible for manufacturing safety infrastructure for artificial intelligence (AI) startups.
Moonbounce recently shared news that it closed a $12 million funding round, money that will be central to broadening their service reach. The company is focused on improving safety resources in the air, on the ground and online. They will add another layer of security wherever content is produced. Levenson joined with Ash Bhardwaj to co-found Moonbounce. Bhardwaj, Levenson’s former colleague at Apple, was influential in engineering the massive cloud and AI infrastructure that supports the tech giant’s core products.
Levenson’s inspiration for Moonbounce came from the challenges he faced while developing these content moderation projects from inside Facebook. He recalls the insurmountable challenges that human reviewers faced. In preparation to testify, they had to learn a complicated 40-page policy guide, translated into more than six languages. This experience led him to advocate for “policy as code.” He is an advocate for automating accountability mechanisms and integrating them directly into the algorithmic black boxes that shape content moderation.
Moonbounce serves three primary sectors: platforms that handle user-generated content, AI companies that create characters or companions, and AI image generators. The company is indeed on an unprecedented uphill climb. It analyzes more than 40 million real-time reviews per day and provides messaging to more than 100 million active users per day.
To improve its abilities, Moonbounce has created its own large language model. Our AI model analyzes complex content in real-time and serves up answers in less than 300 milliseconds. This quick turnaround time is especially important in balancing content moderation with both effectiveness and efficiency.
We’re a third party sitting between the user and the chatbot, so our system isn’t inundated with context the way the chat itself is,” Levenson stated. This divide lets Moonbounce operate at a higher level of maturity, assuring safety without being bogged down by the minutiae of anticipated and unanticipated user behavior.
To this end, AI companies are starting to see the value in soliciting support from third parties to enhance their safety infrastructures. “Safety can actually be a product benefit,” Levenson noted, emphasizing that enhanced safety measures can improve user experiences rather than detract from them.
Moonbounce has invented technology that can reduce the speed of content distribution while allowing sufficient time for a human to review. It doesn’t just filter risky material by analyzing material in advance. This flexibility is extremely powerful. It stops toxic material from circulating among users while allowing more robust investigations to occur.
Reflecting on the challenges faced by human moderators, Levenson remarked, “It was kind of like flipping a coin, whether the human reviewers could actually address policies correctly, and this was many days after the harm had already occurred anyway.” This recognition highlights the demand for a more automated and efficient solution in today’s rapid digital environment.
Levenson’s commitment to improving safety goes beyond checkbox compliance. That’s how he envisions a future in which AI applications start to incorporate guardrails to create more beneficial experiences for users. “We hope to be able to add to our actions toolkit the ability to steer the chatbot in a better direction to essentially take the user’s prompt and modify it,” he explained. His vision is to change the way chatbots work from passive listeners to proactive partners that enhance productivity and engagement.
Lenny Pruss, general partner at Amplify Partners and investor in Moonbounce, was highly confident in the company’s vision. “We invested in Moonbounce because we envision a world where objective, real-time guardrails become the enabling backbone of every AI-mediated application,” he stated.
Leave a Reply