Web3: The Internet of AI

Introduction

Web3 is viewed by some as the next phase of the Internet, and it’s hard to argue that it doesn’t offer benefits that could lead to huge changes in the status quo. When viewed as a standalone development, Web3 fixes many of the problems surrounding data concerns while also providing a workable infrastructure for the storage and ownership of digital assets. But I think it offers so much more than that. To me it’s an online framework that feels custom-made for the most important tech development of our age. In short, Web3 is an ideal playground for future generations of AI.

Paradigm Shifts

It seems like everything’s shifting right now, and I think that’s because we are in the epicentre of a systemic reset that’s impacting most areas of life. In large part it’s technology that has driven us to this point – the digital interconnection of communities across the world, and the underlying technologies that now drive society have changed how we live in irreversible ways. Moreover, the systems we have become used to are adjusting as technology evolves, and as a result we find ourselves at the end of the old world and the start of the new.

One key area of change that I don’t think is getting anywhere near enough attention is that which is occurring below the surface of the Internet. Deep underneath the veneer we interact with in our day-to-day lives (apps & websites), a wave of alterations to the mechanisms that drive the web is taking place, and too few people realise it’s happening. I’m not surprised the majority are oblivious though – at present the technological shifts occurring in this space are of a fairly technical nature. However, when distilled down to the simplest terms, these changes are easy enough to comprehend - we are seeing a reversal to the established ways in which information is transmitted, stored and accessed online. 

That sounds incredibly dull and uneventful I know but believe me when I say that these changes are going to impact how we interact with the digital and physical world in significant ways. And it’s only a matter of years before everyone's aware of how markedly different a Web3 world will be . . .

Web Evolution

For those of you who are unfamiliar with the term, Web3 describes a new evolution of the Internet. Currently, 99% of what we interact with online would fall under the bracket of Web2 – essentially dynamic websites: cloud-based platforms, streaming and social media. And for anyone who is already familiar with Web3 but sceptical about its deep association with cryptocurrencies – I hear you. Many people see the crypto market as a speculative mess, filled with Ponzi schemes, dodgy dealings and worthless coins. While that’s partially true, beyond the noise created by less reputable elements of this space is an infrastructure-in-progress that I believe will play an integral role in all of our futures.

If you’re reading this thinking ‘ok, so Web3 is clearly a thing, but I’m not entirely sure what Web1 & Web2 are?’ you’re not alone. Few people chart the evolution of the Internet and instead choose simply to use it. And why not - that’s what it’s there for. 

But in order to understand the impact of Web3, it’s important to first have a grounding in how the Internet has evolved . . .

Web1 was what came about first – static websites that were basically just glorified digital books at their heart. Those old enough to remember the emergence of the Internet in the mid 1990s will just about remember a time of clicking through pages that allowed for little or no interaction. This was Web1. 

Web2 evolved from the first wave of static websites, leading to interactive forums, video streaming and social media platforms such as Facebook. The digital world we know today - from work to entertainment - is built on Web2. But that’s likely to change in the not-too-distant future.

Web3 is a departure from the infrastructure on which Web1 and Web2 were built. Instead of huge amounts of data being stored centrally by large companies, Web3’s network distributes data across a vast network of computers

Web2: A Flawed Machine

Before I elaborate on the specifics of Web3, it’s a good idea to highlight some of the negative things that Web2 has given us. Because despite a long list of benefits (ease of communication, ‘free’ services, a huge level-up in entertainment and remote working to name but a few) it has come with some drawbacks. 

In the Wild West world of the early Internet, we witnessed a small group of companies fighting for dominance. Internet giants that we now refer to as ‘Big Tech’ (e.g. Google & Meta) are as synonymous with the Internet as the BBC is with television, and often just as trusted. But it’s important to realise that these entities exist only because they were first-movers and innovators who successfully established ownership and monopoly over everybody else. Unlike the BBC, they are not public service providers – Google, Meta & Amazon are behemoth, money-driven companies that have eaten up much of the competition having firmly planted their flags in the digital soil.

So what? That’s fine. They offer us free access to webmail, video streaming, social communication and a range of online apps and benefits, so who cares? But hold on.

Unfortunately, there is nothing free about any of the services Big Tech gives us. Far from it. Google, by example, is one of the most profitable companies on the planet. No organisation reaches a $350 billion annual turnover by giving stuff away gratis. We are paying for their services alright – and the currency is data.

Imagine for a second that data is crude oil. Now imagine that humanity is one gigantic oil field. Well, Big Tech ‘struck data’ in the early 2000’s and have since been drilling it for over two decades. Like oil, data needs refining, but they developed algorithms that are very effective in doing that. At its core, data really is like a natural resource when you examine it, as it provides raw material for wealth generation to whoever owns the land from which its extracted. And guess what – just a tiny number of companies own all of the data mines in the world.

But Web3 aims to change this.

A Decentralised Internet

Put simply, Web3 is a rethinking of the Internet’s architecture. At present, most of our data is stored centrally on ‘the cloud’. What sounds like data swirling above our heads is in fact massive data centres on the ground, run by Big Tech. We tend to view the Internet as a ‘web’ of interconnected devices, which it is. But the bulk of the data we access is siloed into giant data storage facilities managed by central companies. Web3 alters this, by instead distributing data across a large number of computers (known as nodes) owned by individuals or small companies. But the magic is that nobody ‘owns’ the data because no one owns the network. You see, this is a decentralised Internet that operates via a series of DAO’s (Decentralised Autonomous Organisations) where voting and user voice takes the place of traditional company structures. 

So, all that is straightforward enough, right? But what I’ve described so far doesn’t explain how Web3 works. And before I attempt to do that - be warned, this is complex stuff. But I will try to explain both its origins and current state in is as simple a way as possible . . .

Web3 Mechanics

Way back in 2009, a little thing called Bitcoin emerged. Most people have heard of Bitcoin and many know about its mysterious origins. But for those who haven’t - Bitcoin is digital money created by an unknown individual (or individuals) who went under the guise of Satoshi Nakamoto. It was brought into being to act as an answer to the problem of big banks and financial crashes. Bitcoin exists on a network of computers (nodes) known as a blockchain. The nodes form a ‘chain of blocks’ that validate and confirm that information is correct, and each node stores what’s known as a ‘ledger’. Every node stores an identical version of the ledger that says who has how much Bitcoin, and because of this, if a hacker tries to alter the ledger on one of the nodes to say that they own, say, 100 Bitcoin, the other nodes will disagree and revert the ledger on the hacked device back to its previous ‘correct state’. Finally, all of this is made possible through the wonders of cryptography – the secret sauce that makes everything secure and immutable. That, in essence, is how Bitcoin works, and further - it's how the whole of Web3 works. Web3 is essentially a series of many blockchains that shifts Internet usage from centralised entities, over to a decentralised network of computers spread across the globe. 

Although Bitcoin wasn’t created to kickstart a new version of the Internet, it did. People quickly realised that the same structure – many computers storing and verifying information to keep things in check - could be used not only to distribute digital money, but also to transfer information and digital assets (images, video, sound etc.) through what is known as smart contracts. Ethereum was the first network to popularise smart contracts, and many others have sprung up since -Solana, Cardano and Avalanche to name but a few. Therefore, many thousands of networks now exist with the purpose of tackling various use cases – file storage, distributed computer power, financial assets and so on. So, this is the structure of an emerging, new Internet, and yet few people realise it even exists. Or at least most people have heard of bits of it (Bitcoin & NFT's mainly) but don’t realise that it represents an alternative to the web they're used to.

This is because currently Web3 is very much a background network that operates with fairly clunky interfaces. It’s in a nascent state – new, fledgling and underdeveloped. One thing that’s important to underline is that it’s driven and funded by digital assets – crypto (that 'big scam’ you occasionally hear about in the mainstream press). It’s unfortunate that crypto has developed this reputation, but it’s one that’s rightly deserved in some cases. Currently, Web3 is in its own Wild West stage of development, much like Web2 was back in the early 2000s. As a result there are people keen to exploit this emerging landscape with get rich quick schemes – mostly ‘meme coins’, which you have probably heard about. But at its heart, Web3 and the crypto that drives it, is fundamentally sound technology and many predict it will become a major aspect (if not a key aspect) of the Internet infrastructure of the future.

But hang on – this article is called Web3: The Internet of AI, so how does all this relate to Artificial Intelligence? Allow me to explain.

The Era of Agentic AI

At time of writing the AI we use is incredible but limited. Currently, if you want AI to perform a task, you need to provide prompts to guide it toward the desired outcome. Don’t get me wrong – this is magic stuff, and it still blows my mind that it exists. But it isn’t autonomous. It can’t carry out a series of tasks on its own initiative. That’s where Agentic AI comes in.

Agentic AI isn’t really with us yet. Or at least, versions are available but they are not widely used at present, and don’t operate in a way that makes them universally useful. But behind the scenes, development and testing is well underway with OpenAI and others gearing up to release AI Agents into the world in the near future. When they do, and if they live up to expectations, the impact on life, work and the way we interact will be staggering.

That’s because AI Agents are autonomous versions of AI – armies of individual ‘workers’ who will be able to go off on their own to do our bidding. Want a holiday booked? AI Agents can do that. Need a website and online store created for your business? Agents can do that too. Not only will they likely have amazing independent autonomous abilities, but they will also be able to interact with each other. This will create a substrate of digital workers – a virtual workforce that can do most of the things we do when interacting with computers and the Internet. Worried about your job security? You should be.

Intersection of Web3 & AI

But what does all of this have to do with Web3? How do AI Agents and a blockchain-driven Internet intersect? Well, for one - Web3 is a perfect environment for AI Agents to ‘inhabit’. I don’t mean that they will physically live on the various nodes of a blockchain-based Internet. That’s not how digital entities exist, and not requiring a ‘home’ is just one of the many strengths AI has over us, real-world human workers. What I mean is that the network of blockchains that make up Web3 are a great framework for them to operate in. Web3 is decentralised, efficient and capable of facilitating the interactions and transactions that AI Agents will need to execute. 

When Agentic AI first emerges, it will exist in models born from centralised organisations. OpenAI’s ChatGPT, Google’s Gemini and X’s Groq are run on Web2: the Internet controlled by Big Tech. But once the formula for these models is out there, decentralised AI Agents are bound to be developed. Purpose-built Web3 blockchains such as Fetch.AI exist ready and waiting for when this takes place. So, while Web2 will be the mother of Agentic AI, Web3 might end up being the playground of its offspring.

Who Owns AI?

One question surrounding AI today is who actually owns it. Right now, AI development is almost entirely controlled by Big Tech - OpenAI, Google, X, and a handful of other corporations. These companies are gatekeepers to access, determining who gets to use AI and under what terms. Currently, AI is often a rented service rather than a personally owned tool, but what if it wasn’t monopolised in this way? What if instead of being managed by corporations, individuals or decentralised networks could own and train AI?

This is where Web3 could offer a very different scenario. Instead of AI models being locked behind corporate barriers, decentralised platforms like Fetch.AI and SingularityNET are pioneering open AI ecosystems, in which models are trained and deployed on blockchains. In these cases, AI agents wouldn’t be tied to a single company’s infrastructure but would instead exist within Web3 marketplaces. This would allow individuals and businesses to own, train, and rent out AI services without centralised control.

And so, the intersection of Web3 and AI raises some promising possibilities. For example - could it enable each of us to have our own personal AI? If we could, it would mean that AI isn’t just a tool we access, but something we can control & trust. A future such as this is why I believe Web3 is a natural home for AI. Not just because it offers a digital landscape to operate in, but because it could improve our freedom in an AI-driven world. 

If AI agents are personally owned rather than rented from Big Tech, this could also lead to a new digital economy where individuals profit from AI instead of centralised organisations. If this takes shape, it could enable a future where agents are leased out as digital workers, generating income for their owners in a decentralised model of AI-driven industry.

Regulatory Hurdles

It's wise to reflect on the inevitable impact of regulations that governments are working to put together right now. These frameworks are being developed to oversee both Web3 & AI. While Web3 clampdowns focus largely on crypto due to its impact as an economic disrupter, AI regulation is seeking to address very real issues of a technology that poses as many dangers as it does benefits. While government interventions in both fields will shape aspects of how they operate, I don’t believe that any regulation can push Web3 or AI too far off-course. As both are simply code, spread across the expanse of Internet, any controls imposed on either are unlikely to prevent their progression to a notable degree.

Messy UX & Crypto

Currently, Web3 is complex and inaccessible to most users, because it lacks a refined and intuitive interface. Instead, current users are forced to navigate a complex system of crypto wallets and sites that just aren’t designed for the average Joe. I’m a bit techier than most, but I struggled to get my head around connecting wallets, storing assets and ensuring strong security practices when I first started exploring Web3. 

But AI agents could solve this problem. Serving as an intelligent go-between, Agentic AI would be able to automate blockchain interactions and eliminate the need for clunky, manual navigation. In doing so, agents could make Web3 feel just as intuitive as Web2 without sacrificing the benefits of decentralisation.

The need for AI Agents to transact on our behalf (e.g. buy stuff on Web3) is the factor that I think makes Web3 an ideal environment for their use. As Web3 is driven by thousands of digital currencies - each assigned to different services, through blockchains - AI Agents will be able to use them to seamlessly transact on our behalf. If AI Agents can act as our intermediaries in this way, they could be a significant catalyst in moving online activity from Web2 to Web3, allowing users to opt for decentralised services instead of Big Tech's offerings.

Currently, the complicated interfaces of Web3 are a massive barrier to mainstream adoption. In its current state it’s like a car engine without a chassis – full of potential to move us forward, but not capable of getting us there. Agentic AI could solve this problem, and I think once it’s out there, this will happen organically.

The AI Interface

So, Agentic AI may well become the user interface of Web3. However, the way we communicate with AI right now isn’t perfect – detailed text-based instructions or voice-commands are needed to get optimal results. And even then, current AI can only create things in-App or provide an augmentation of our thinking. We need to act upon AI’s guidance and expertise, or set up complex automations to get the results we want. But both AI and Web3 (in their current forms) are at early stages in their development. And the lack of truly productive AI and a usable decentralised Internet is one that I believe AI Agents will solve.

Wearables seem to be the most likely devices we will use to communicate with AI Agents in the coming years, which will act as a proxy through which we can interact with Web3. AR (Augmented Reality) glasses that blend elements of the digital world with the physical are the most obvious candidates for this role, replacing more labour-intensive typing and swiping. These devices will enable us to direct AI via simple voice-commands and hand-gestures, requesting services and getting results with a notable reduction in effort and steps. 

AI memory (such as that seen in ChatGPT and Google’s Gemini) is already very effective in enabling tailored output, using information they have learned about us through previous interactions to refine their responses.  But imagine how this could amplify our interactions with AI Agents! When paired with a detailed, consistent memory of us, Agentic AI will take this to a whole new level, allowing agents to second guess missing details based on data they have gathered over time. This will - for example - allow AI to seamlessly book a holiday based on our tastes, interests, and budget with just a short voice command. And that’s nothing compared to what this technology could provide in every area of our lives, let alone businesses.

The Internet of AI

To conclude, I predict that AI Agents could bridge the divide between us and Web3, translating the current mess of blockchain navigation into a workable and usable system. Further - I foresee massive benefits in their symbiosis and a lot of potential in their combined use. When Agentic AI and Web3 finally intersect, the Internet itself could evolve into something fundamentally new. The collision of these two technologies can work to safeguard our data, provide personalised AI, enable easy digital ownership and shift a significant proportion of online activity from Web2 to Web3. That’s why I think that one day Web3 will be viewed as The Internet of AI.