Adding AI to Web3 will spark a Cambrian explosion of innovation — but not anytime soon
Jack O’Holleran
6 min read
Imagine your favorite social media platform started using a new AI bot detection tool, and for some reason, your account kept getting flagged as fraudulent despite you being a real, human user.
You, and anyone else mistakenly getting flagged, would have little recourse today.
With millions, even billions, of users, it’s almost impossible to get noticed by customer service reps on some of the larger platforms. And if you wanted to get the platform’s algorithm to consider additional data points, such as metrics that would prove the humanity of you and others in your situation? Yeah, good luck.
But what if the platform’s artificial intelligence model was integrated with the blockchain?
The factors that drive the model’s bot determinations would be publicly available on chain, for anyone with an internet connection to see. The AI model’s decision framework would be transparent, and if it were tied to a blockchain-based decentralized autonomous organization (DAO), members of the platform could make a proposal for how to change the model so that it doesn’t incorrectly label people as bots.
There are countless other things one could vote on, of course — everything from content moderation standards to user experience decisions. The broader point? Fully integrating AI models with Web3 technology can unlock greater transparency, greater value exchange, greater decentralization, greater education, learning and communication.
That promise has people all over the Web3 ecosystem raving, to the point that their shared excitement over AI and Web3 has become easily memed. And while that excitement is valid, let’s dump some cold water on this whole thing: We’re still probably a decade away from seeing true AI-Web3 integration become a reality.
The current blockchain AI market, valued at US$230 million in 2021, is expected to grow into a billion-dollar industry within the next decade. It could potentially get to that valuation much sooner — but it will have to first overcome the fact that decentralizing AI is a difficult and costly affair.
Doing the millions, even billions of transactions required to run an AI model is already an extremely expensive affair, and doing so on the blockchain is significantly more so. That output will require much more from smart chips than is currently possible, similar in many ways to the massive advances that will be needed to power another high-transaction Web3 innovation: the metaverse.
AI-empowered blockchains and protocols could stack the benefits of machine learning with the decentralization and aligned incentivization of Web3. That stacking can lead to exponential gains, optimizing not just work through AI, but also the way the value from that work is distributed through the incentivization, ownership and transparency models enabled by Web3 technology.
Powered by AI, here are five Web3 use cases we’re likely to see in the future:
DeFi with AI-boosted risk assessment: AI can significantly enhance decentralized finance applications by providing advanced risk assessment models that evaluate the creditworthiness of a user requesting a loan or determine the risk of an investment product. Since the blockchain ensures transparent and immutable record-keeping, AI models can leverage this data to make more accurate predictions.
AI-driven NFTs: As NFTs evolve from static to dynamic entities, AI can play a significant role. For example, AI could enable the creation of “smart” NFTs that change over time based on certain conditions or inputs. This could lead to a wide range of innovative applications, such as NFTs that adapt their appearance according to the time of day or an artist’s mood, or NFT-backed virtual characters that evolve based on user interaction.
DAOs managed by AI: Decentralized autonomous organizations can leverage AI to automate decision-making processes and improve the efficiency of operations. For instance, AI could help with optimizing resource allocation, making predictions about future trends, or even voting on proposals based on predefined criteria. The parameters guiding these AI models could be set and adjusted by the community, providing a balance between autonomy and human oversight.
Personal data monetization: Web3 gives individuals greater control over their personal data. Combined with AI, users could not only control who has access to their data but also monetize it if they choose to. For example, users could allow AI algorithms to use their personal data to improve their models, and in return, they could receive compensation in the form of cryptocurrency.
AI-powered metaverses: Artificial intelligence can be integrated into virtual worlds to create more realistic and dynamic experiences. For example, AI could be used to generate unique, real-time content in the metaverse, such as creating personalized quests in a game or simulating realistic weather patterns in a virtual world.
Next-generation blockchain layers will incorporate AI into the core components of their network, expanding efficiency in storage and other essential functions. One can imagine a world where the validator market consists of not just human validators but also AI ones, enhancing security on protocols as well.
Eventually, AI will be incorporated in a way that it can essentially “govern” Web3 blockchains and networks. Instead of a DAO voting on every small tweak or adjustment to the protocol, the AI model could be given wide purview to make decisions that keep the DAO working efficiently.
The community could adjust this purview based on their own values and interests. Importantly, it could also adjust the parameters by which the AI model makes decisions about the network — and due to the transparency of the blockchain, these parameters could be public and easily accessible for all to see.
Right now, it’s difficult for ordinary users — even large communities of them banding together — to compete against massive platforms with huge amounts of technical and financial capital at their disposal. AI’s ability to augment human capabilities could help level the playing field for those ordinary users, combining with DAOs and other Web3 organizations built on the blockchain to better distribute ownership and governance.
This final stage of AI and Web3 integration will be difficult and costly to achieve, which is why it won’t happen overnight. In fact, it will take much longer than much of the hyped-up pieces that are being shared across the internet today.
Still, once that integration does come, it will open up a whole new galaxy of apps and services that reward people with more ownership and control. And the level of innovation that emerges could be orders of magnitude greater than what we can imagine today — akin to humanity using flip phones in 2005 without realizing that in a decade they would be able to click a few buttons and instantly call drivers to their location, order groceries, code applications and do countless of other previously unimaginable things.
