The rapid expansion of AI systems has ignited a fundamental question: How do humans stay meaningfully involved as intelligent software grows more capable, autonomous, and ubiquitous?
The second episode of Exponential: A Nexus Podcast features Jordan Gray, co-founder of PublicAI, a startup working to ensure human participation remains central to the future of artificial intelligence.
In a wide-ranging discussion, Gray outlines how PublicAI is developing infrastructure to support distributed human intelligence — a system designed to complement large-scale AI with real-time human oversight, input, and accountability.
“We want to be able to pay people fairly for their participation in these systems and take a proactive stance as the AI work revolution happens.”
Beyond the hype: Why inference matters
Much of the AI conversation today centers on model training — data pipelines, fine-tuning, and emerging risks like model collapse or hallucination. But Gray emphasizes that real-world value is often captured not during training, but in inference: the ongoing process of making predictions and decisions based on trained models.
This is where AI meets business. Whether it’s workflow automation or real-time decision support, inference is where human feedback becomes essential for quality control, reliability, and ethical alignment.
“The AI itself can’t be held accountable for decisions… ultimately a human is still in control when it comes to important, impactful decisions.”
A global network of human oversight
PublicAI’s core thesis is that humans must remain embedded in every stage of the AI lifecycle. To that end, the company has built a network of over 1.6 million contributors who annotate data, validate model outputs, and provide structured feedback—all coordinated through blockchain primitives.
“If you want to do distributed human intelligence, you need a way of paying people that works globally — and web3 is perfect for that.”
An example: capturing high-quality photos of dog noses to help train biometric identification for pets. Contributors passed a subject-specific quiz, submitted original images from multiple angles, and had their data verified both mechanically and through peer review — earning stablecoins in the process. The result was a clean, trusted dataset delivered to the client within days.
Distributed systems, real incentives
What makes PublicAI’s model particularly compelling is the blend of rigorous coordination with aligned economic incentives. Reputation systems, staking mechanisms, and tiered roles (from contributors to validators to judges) create a governance structure designed to reward authentic participation and discourage manipulation.
“When you’re talking about decentralized coordination… what you build winds up looking a lot like governance.”
This approach resonates with broader trends in digital infrastructure, where systems like Nexus are also designed to scale globally while preserving accessibility. Together, projects like PublicAI and Nexus represent a new class of internet-native institutions — coordinated by computation, powered by people.
“It dovetails perfectly: Nexus bringing the distributed compute, PublicAI bringing the distributed intelligence — together building a more human, happier vision of the future.”
Building toward a participatory future
The conversation closes with a forward-looking reflection on what’s at stake.
Without systems to value and compensate human input, there’s a risk that AI’s future could become exclusionary or extractive. But by taking tangible steps now — through architecture, economics, and real deployment — teams like PublicAI are showing what an alternative might look like.
“Unless somebody is taking tangible steps toward building a future where human participation is valued, it’s very easy to spiral into passive dystopian outcomes.”
This episode of Exponential makes one thing clear: the future of AI isn’t just about faster models. It’s about designing systems that remain accountable, adaptable, and deeply human at their core.
You can find Exponential on Spotify, Apple Podcasts, and YouTube. Whether you’re a builder, investor, or simply someone fascinated by the future of computation, we hope you’ll subscribe and join the conversation.
