The Real Value of AI in Crypto — A New Paradigm for Web3

Published in

Coinmonks

8 min readMay 18, 2023

Just a few weeks ago I told myself if I heard of another AI token project or “AI and Web3” project I would leave that space immediately. I wasn’t seeing the value, didn’t look hard enough to find real use cases and the so-called AI revolution didn’t hit me. Then I heard this Bankless Pod. My mind was blown. So for anyone struggling to discern which projects in crypto are leveraging AI properly, this is for you. The real and fundamental bull case for AI and Crypto (in layman’s words).

Bill Gates said in a recent blog there were two demonstrations that struck him as revolutionary in his entire lifetime. First was the graphical user interface that we know so well developed into Windows, the backbone of Microsoft. And second was OpenAI. In the 67 years of this named man’s life, if he believes the “Age of AI” has come, we’d probably best give it some credit. So, acknowledging that AI has the potential to resolve humanity’s most pressing concerns of inequality in education, health and climate change, all the while at a micro level it boosts productivity and efficiency at unprecedented speed, then what does this tremendous leap forward in humanity mean for crypto, and how does it leverage blockchain?
One step further, do AI and crypto necessarily need to converge?

Let’s dive into what crypto means for AI, what AI means for crypto, and where the investment opportunities are in its convergence.
(Of course, nothing is financial advice.)

TL;DR

Crypto is the only solution to AI’s greatest pain points — autonomy, privacy, and authenticity verification.
AI opens a new paradigm for Web3 and Crypto, as AI agents are also target users in permissionless and decentralized systems.
AI x Crypto investment opportunities for public market participants are not easy to find since AI tokens are mostly narrative-backed. The less-than-obvious opportunities are introduced below.

Crypto is the Only Method for AI Check and Balance

As concerns over AI development have been rising and the threats of uncontrolled AI call on an existential crisis for humanity, we know AI influences reach beyond anything we have ever imagined. AI, is simply, overwhelming humanity. Especially as large language models (LLMs) have begun producing quality content, perfectly customized and deep fakes are harder to discern, the power and influence of AI is unpredictable. Immediately, how the next US election will play out in this swamp of false information or fake news driven by unverifiable AI models will be interesting. That said, crypto is the only technology that can steer AI. Only crypto can provide permissionless payment and execution rails which is critical for AI, and crypto is the only channel by which AI’s greatest threat to authenticity and privacy can be addressed.

Let’s dive.

Issue 1: AI Needs a Supportive Interface, Payment and Execution Rails

Traditional (Web1,2) platforms are allergic to AI. Meanwhile, blockchain provides a supportive interface for AI. Considering bots nowadays are different from traditional scams and can produce helpful and quality content, in order to flourish, AI needs an open-source, permissionless system that allows open access to not only data but payment and execution rails.

The only space that allows this is crypto, leveraging blockchain.

In crypto, AI can do anything that a person could on every.single.level. Payments? Use any wallet of your taste. No proof of personhood is needed. Need execution? Welcome, code is law with smart contracts.

Let me explain.

There is an AI agent, for understanding’s sake, with the name ChatGPT. To say it is an AI agent, means it can communicate with other AI models, interacting at the backend (Baby AGI, AutoGPT). Simply said, your ChatGPT can work with other ChatGPTs, respond and execute commands to and from each other. This can extend beyond just one AI agent to multiple AI agents.

Let’s say you give a high-level command such as “create the best possible innovative but niche website for social interactions based on current social trends”. Your ChaptGPT could then hire community managers to moderate any in-person interactions, create a curriculum, and schedule speaker sessions, all the while finish building the website, the brand, and develop a marketing strategy. Since these bots are autonomous, they have the capacity to hire, pay and trade just as an in-real-life operations employee would. But in the Web2 world, AI’s capacity to execute would be limited as most activities require KYC or some sort of proof-of-personhood. But with crypto, all of these barriers are lifted. In the crypto landscape, AI bots have access to anything a human does — it can leverage crypto’s permissionless, decentralized and open-source culture. Imagine it wanted to scale itself, with crypto, it has decentralized compute to build on!

AI can essentially begin creating its own world, with its own species, creating a value chain or ecosystem made and run entirely by themselves (AI agents). But since AI is incapable of creating its own high-level goals (at this point in time) people will use AI agents as a tool and marketplaces where AI agents can interact with each other to achieve whatever goal they have been given, will be something to look out for.

This was a simple thought exercise (quite scary if you think about it, especially considering this became possible just a month ago) but it applies to any business opportunity including decentralized finance, social networks and gaming. The competition for maximum efficiency, productivity and optimization will simmer down to strategy rather than execution, for the league itself will be run by AI. The point here is that for anything of this to be truly executed, crypto is the only solution for AI to reach its full potential and autonomy.

Issue 2: AI’s Privacy Concerns

The second issue is that AI is quite transparent… aka no privacy. But AI only becomes useful when it has a vast database fed into it. Most AI models today essentially collect and combine all the information on the internet, store your data, and use it again as a part of its training set. In this process, the more you engage with the model your private data is at risk (alongside many other privacy risks). But in order to leverage the potential of AI, it needs data that is extremely private, so what can we do?

Crypto has the unique solution to this, zero-knowledge proofs (zkp).

Let’s say, you want to train an AI model to diagnose cancer or find a solution to a unique tropical disease. A few hundred patients are willing to contribute by disclosing their private medical records such as X-rays but are uncomfortable with sharing any personal details that could later identify them as a participant.

Through zkp, data can be contributed while maintaining anonymity. Digital signatures from the doctor can verify the legitimacy of the request and information, and the necessary medical data such as X-rays or blood levels from the contributor can be transferred without disclosing any other private information such as their social security number or address thus protecting their privacy.

Read more about how it works here: Source

Issue 3: AI’s Authenticity and Verification Concerns

Not only that, in this world of disinformation, fake Drake songs and AI influencers, how do we know what is a deep fake and how do we verify sources? How do I discern fake news from real, if photos, videos, voice, tone and manner of speech can be perfectly imitated and created flawlessly?

It’s easy with crypto. Simply verify authenticity with a digital signature.

This notion was mentioned briefly in privacy protection above, but using cryptography, public keys and private keys can never be replicated by a bot, nor imitated by AI. They are yours, and yours only. Sign off your original piece of content with your unique key and anyone who recognizes your key can verify its authenticity.

When digital signatures meet zkp, the magic amplifies. You can verify any subset (a clip or edit, such as resolution change, colour tone change etc.) of the original as long as you publish the chain of custody.

Aside: Something to remember here is that digital signatures can prove authenticity but it cannot necessarily verify personhood. Digital signatures prove you are looking at the original version from the original source. But at the deepest level, in order to prove personhood, in-person interaction would be inevitable.

AI will be Crypto’s Largest Target User — Investment Opportunities Found in this New World

As mentioned in this article that was the inspiration for this medium, if blockchain can be understood as a network, in 10–20yrs AI will “dominate the edges of networks”. Humans will continue to develop the thinking and provide direction to what must be done, and how rules should be understood, but the players and participants will not all have to be human. The lines will be blurred. Then how can we reap the economic benefits of this innovation?

Marketplace for AI agents
Perhaps the most obvious but hard to find. Look out for projects that allow AI agents to interact with each other and leverage blockchain and crypto customs.
Coins AI agents can use
The most simple option. Buy coins that AI agents will use.
Projects that have low barriers for AI integration
For founders or investors with access to private markets, look for projects that are AI-friendly but not AI-reliant. Most AI-reliant projects are probably leveraging AI for more novel purposes or are just scratching the tip of the iceberg. These projects may not reap long-term returns. Remember in the crypto world, AI should also be considered a user.
Traditional — Chips, AI-centric opportunities
Jump on the traditional AI infrastructure investment opportunity train. I’m not going to pretend I know or understand it, but there must be things AI relies on for computing power, storage etc. DYOR

In Conclusion…

The argument of how crypto stands for decentralization and AI for centralization is outdated. The innovations of these early technologies have opened a new paradigm for the world, and by leveraging each other’s characteristics new opportunities will rise from their synergy alongside new challenges and complications. There are more exciting reasons and influences of AI and crypto’s convergence like decentralized compute that this medium did not cover. Refer to the following resources to dive deeper into this mind-boggling revolution and have fun exploring.
I’ll be back with another curiosity.