Generative AI Inside Programmable NFTs Is New — and Can Accelerate the Sector Beyond Hype
In their infancy, NFTs have suffered mightily from both misconceptions about the technology behind them and manipulation by fraudsters and hackers. Issues have arisen because many people, including creators, rushed to treat them as monetized cryptocurrency projects, tradable assets, and get-rich-fast investments. The more thoughtful, functional opportunities possible with them have often been overshadowed by hype, speculation, speed, and greed.
That’s finally starting to change — and it isn’t because brands made them collectibles or high-end coupons. It’s because evolving technical capabilities are steadily improving NFT functionality.
It’s still true that an NFT refers to an asset, most often an image, tokenized with a unique digital signature recorded on a decentralized, immutable blockchain like the Ethereum network. This signature on a distributed blockchain ledger, across thousands of computers worldwide, ensures that the NFT’s existence and transfer can be verified with an extremely high level of certainty — without relying on a central authority like a bank or broker. That’s well and good, but the public has largely remained unable to see this in practice, translate what it means, or find relevant applications to their lives and work.
With the emergence of programmable NFTs in 2023, we can now directly embed content and executable programs inside an NFT, which creates a significant difference in how we experience them, their utility, and even the jobs they can perform for us. My team’s new Immutable Miniverse Format (IMF), for example, immutably encodes — and can encrypt — tailored messages, games, music, binaries, and interactive experiences within an NFT, visualized as a string of secret code in motion. Importantly, this self-contained “miniverse” is embedded as part of the NFT artwork — without the technology burden of maintaining Web3 access points when being used on traditional social media.
The Next Level with Generative AI
Fusing NFT art and programming code in a self-contained way is a significant development.
Programmable NFTs can be leveraged for work and play, shared on today’s social media platforms (Web2) and even serve as the building blocks of a kind of “app store” for Web3 and a more decentralized economy, comparable to today’s mobile phone apps. Additionally, the use cases in everyone’s mind involve generative AI technologies.
For good reason: the ultimate NFT utility may be the capacity to tokenize intelligence in a decentralized and immutable fashion.
Here, we’re not talking about using generative AI to churn out NFT art, rather the focus is on integrating LLM chat capabilities into programmable NFTs and ultimately empowering owners to program and mint their own NFTs to do this. My team is enabling NFT holders to have a private conversation with and within their NFT about the domain and knowledge embedded inside their NFT’s miniverse. This means leveraging a customized, fine-tuned version of an LLM, while not subjecting any input to third-party services that could opaquely use the data or violate privacy — a criticism leveled at OpenAI and others in a typical centralized GPT setting.
In fact, if you’re maximizing the strengths of both technologies, NFTs and gen AI, then:
- The embedded information on which an LLM conversation is based should be immutable, individualized, and decentralized.
- The conversation with such embedded information must preserve privacy. It does not need to reference any public GPT services.
- NFT creators must be committed to decentralized AI on Web3. Future iterations should and will allow a user to run their LLM model locally on the user’s own computer.
The protection of privacy possible with NFTs is a differentiator compared with using generative AI in a web application. A criticism of ChatGPT, BARD, and similar systems is that input from a user becomes data consumed by the generative AI, including proprietary input, and that can ultimately become available to other users. A generative tool like Microsoft’s Copilot stores input data, while assurances are made that it isn’t used for training large language models.
Still, the concern is that the power of massive data and its control becomes centralized even more by the hyperscaler giants of computing and the few smaller companies they back.
Blockchain’s fundamental decentralization and distribution means that generative AI NFTs can offer an alternative to that system of big data control.
Additionally, AI, as with mobile now, can leverage edge computing, not centralized cloud services.
Long-term Uses of Programmable NFTs in Life and Business
The viability of a technology depends on its ability to solve real problems and to create new possibilities and experiences beneficial to humanity and the world. Right now, we’re seeing a tremendous blooming of new uses for generative AI. When sharing information or when interaction needs to be secure, private, and impervious to short- or long-term systems failure or third-party dependency, the integration of NFTs and AI will prove especially handy.
In the personal realm, consider how you might preserve and interact with family history and important private documentation shared among family members over the long term, without pieces of that being scattered or permanently lost. Some families have members who live across the world and with whom they want to share information. Most people have family members dear to them who have passed away or will do so. Imagine an interactive record of your grandparents’ lives — one that you can have a conversation with. Imagine documenting family events from yesterday, five years ago, or fifty years ago that are less likely to be lost to time in a decentralized system. With programmable, AI-capable NFTs, you and your specified family members can own the assets with the data, the knowledge, the documents, and the interactive capabilities, as long as you have an internet connection. Digital existence and private access are distributed and not dependent on permission granted by a central authority or third party.
The same principles hold true for business uses.
Consider proprietary documentation, blueprints, and even trade secrets that need to be shared among specific engineers across a global workforce or from an older generation to a newer one within a company.
With an NFT, you could securely encrypt documentation, share it with an intended party, and, with generative AI capabilities, enable them to ask questions of it — even, in a sense, converse with engineers retired or long gone. Questions and answers about the documentation can instantly be translated into other languages.
Because NFTs are immutable and recorded permanently in a distributed public registry, impervious to deletion or third-party tampering, their potential for authenticating products and verifying lineage in the supply chain has often been noted.
Adding a layer of interactive inquiry with embedded generative AI makes this business use case even more exciting.
A programmable NFT or immutable “miniverse” with generative AI functionality is a new frontier in interactive, customized applications for personal and business use that truly puts ownership in individual hands. It’s time to lift NFTs out of the abyss of financial speculation and esoteric understanding and put the spotlight on their technical capabilities — which can make them useful for everyone.
[To share your insights with us, please write to firstname.lastname@example.org]