We recently had the opportunity to sit down with industry leader Zac Cheah, Founder of Pundi AI, a decentralized platform aiming to democratize the AI data ecosystem. We discussed how the company is bridging AI and blockchain, the role of decentralized data in training AI models, and what the future holds for AI agents.
Below is our discussion.
Good morning Zac, it’s a pleasure to have you with us here. Let’s start with the basics of what AI agents are and why so many people are talking about them in recent months?
Thanks for having me! So, an AI Agent is basically a term that describes an autonomous, goal-driven AI program that can perform tasks or make decisions with minimal human input. Think of them as digital assistants that don’t just wait for commands, but can proactively analyze situations and act on their own.
In this regard, we can see that just a few years ago, AI’s potential was directly correlated to its chatbot-centric capabilities but with the evolution of these agents, we now have entities that can learn from each interaction and adapt in real time.
This is a big shift, in my opinion, because instead of just responding to questions, AI agents can now handle complex workflows and even improve their performance as they go, almost like virtual team members rather than just tools.
Thanks, so from your perspective, what sort of a transformative impact do you see these agents having on different industries as well as on our daily lives?
The potential of these offerings is truly staggering! We’re talking about change on the scale of how the internet or mobile phones did for us. Why? Because these AI agents are capable of adapting to just about any domain or situation they are placed within.
For instance, in the context of healthcare delivery, they can be deployed to sift through medical data and help doctors find actionable insights. Similarly, when it comes to finance, they can be used to run complex analyses or even make routine decisions automatically. Even when it comes to our everyday lives, we could soon be using personal AI assistants to coordinate our schedules, manage our smart homes, or even find the best financial opportunities for us.
Your company Pundi AI seems to be standing right at the intersection of AI and the blockchain. Could you please talk about how you and your team are looking to bring these seemingly disparate realms together?
Our vision here at Pundi AI is about democratizing AI, which means that we want this yet nascent technology to be accessible to as many people as possible rather than being controlled by a few big players — which is unfortunately the case right now. We want to see Pundi AI become a foundry for AI data, providing users with the infrastructure they need for collecting, curating, and sharing high quality AI training data.
That leads me perfectly to my next question Zac, why do we need to worry about the quality/unbiased-ness of the data needed to train these agents?
Simply put, high-quality data is the fuel that drives any LLM and there’s ample evidence to suggest that there is and will continue to be a huge demand for diverse, well-labeled data in the market today. In the context of Pundi AI, we use blockchain to decentralize and crowdsource the data annotation process — to ensure that every byte of information is recorded immutably and that anyone who contributes to the ecosystem can be rewarded fairly and have ownership in the process, which wasn’t really possible before.
Moving on. Can we please talk a bit about some of Pundi AI’s key offerings?
We’ve developed a suite of tools that all work together in conjunction with one another. These include the Pundi AI Data Platform, the Pundi AIFX Omnilayer blockchain, and soon, the Pundi AI Data Marketplace.
In brief, our data platform allows anyone to contribute by tagging or labeling data which can then later be used to train AI models. Whether it’s text, images, or videos, users around the world can log on and perform tasks via our “tag-to-earn” model. Similarly, Pundi AIFX can be thought of as the blockchain backbone of our ecosystem where every action gets recorded on-chain.
Lastly, our marketplace, which is slated to launch later this year, will be like an open bazaar where people can buy, sell, or trade datasets (whether raw unprocessed data or fully labeled, AI-ready data).
Looking ahead, what do you think is next for AI agents as well as Pundi AI? How do you see this space evolving in the near future?
From the outside looking in, the future for AI agents seems incredibly exciting especially since they have the potential to become increasingly more commonplace in our daily life and business. They’ll likely become more specialized too. Instead of one AI trying to do everything, we might soon be faced with an ecosystem of agents, each trained on data specific to its domain.
Of course, with that growth, there will be a need for frameworks to ensure they behave and learn correctly, which is where transparency and decentralized oversight might play a role. In other words, as AI agents become more powerful, it will become even more important to have systems like decentralized data platforms to keep them honest and well-trained.
As for Pundi AI, I’ve already mentioned earlier how once our marketplace goes live, it’s going to validate a lot of what we’ve been building. Imagine being able to spin up a custom AI agent for your business or project, trained on datasets that you either uploaded or acquired through our marketplace. And then imagine that agent operating on-chain with its own set of rules or even a community (via DAO tokens) governing it.
Disclaimer: This article is provided for informational purposes only. It is not offered or intended to be used as legal, tax, investment, financial, or other advice.
