The Story of Hume AI: Teaching Machines to Care About Human Well-being

Sometimes the most important companies are born not from market opportunity, but from moral necessity. In a recent Category Visionaries episode, Alan Cowen shares how a career studying human emotions led to founding Hume AI, a company racing to make artificial intelligence more humane.

From Academic to Accidental Founder

The path to founding Hume AI wasn’t conventional. “I’m an academic at heart still,” Alan explains, noting that he “was going to be a professor until COVID hit, and the job searches at the time got canceled.” This unexpected turn led him to extend his time at Google, where he witnessed firsthand the rapid advancement of AI capabilities.

During his tenure at Google, Alan helped pioneer the application of data science to human emotion studies. His work culminated in a Nature paper, but he felt constrained by the corporate environment: “It’s tough at a company like Google to really ship things out and be external facing… I felt like we could do a lot more from the technology perspective to make this available to people.”

The Urgency of Now

In 2021, Alan left Google to start Hume AI, driven by a growing concern about AI’s trajectory. His experience had shown him that AI systems, when optimized for narrow objectives without understanding human emotions, could become unintentionally exploitative.

The company’s mission crystallized around a critical insight: current AI safeguards are insufficient. While techniques like reinforcement learning from human feedback help, Alan notes they “can easily be jailbroken or go off the rails as we saw with Big Chat.”

Building the Solution

Hume AI’s approach involves three key components: measure, understand, and improve. The company has built APIs that analyze human expressions across multiple modalities – face, voice, and text. This technology is supported by “massive data sets from around the world where we get people to experience and express emotion using sophisticated psychology experiments.”

The company’s rapid growth suggests they’re addressing a real need. With “over 3500 sign ups” and “over 200 sign ups a week,” Hume AI is quickly becoming a crucial player in the AI ethics landscape.

The Road Ahead

Looking to the future, Alan sees both promise and peril. While AI could enable “incredible applications that we’ll see that will completely blow people’s minds and make the world a much better place,” he also warns of potential dangers, from emotional manipulation to more serious threats.

The race is on to embed human well-being into AI systems before it’s too late. As Alan puts it, “We’re sort of racing to catch up.” For Hume AI, success isn’t just about building a valuable company – it’s about ensuring that as AI becomes more powerful, it remains aligned with human flourishing.

The next few years will be crucial. As AI capabilities continue to advance, Hume AI’s mission to make machines understand and prioritize human well-being becomes increasingly vital. The future they’re building isn’t just about better technology – it’s about ensuring that technology makes the world fundamentally better for humans.