Author: Deep Tide TechFlow
The cryptocurrency market is never short of new concepts.
However, most new concepts are just minor innovations of old tricks; it is precisely these minor innovations that are more likely to bring about new trends and speculation.
The best reflection of this is the method of asset issuance.
From the ICO that started to heat up in 2017, to the subsequent IEO, and now the popular IDO or LBP (Liquidity Bootstrapping Pool)… the start of each wave of changes in asset issuance methods can ignite a batch of new projects and allow some Degen to obtain new profits.
The performance changes, but the core remains the same.
As we enter the year 2024, with AI becoming the "new darling" of the cryptocurrency narrative, issuing assets around AI has become a way to create new concepts.
For example, the recent emergence of "IMO," which translates to "Initial Model Offering."
On March 2nd, a project named Ora Protocol, an AI project, first proposed the concept of "IMO (Initial Model Offering)" on its social media, attracting a lot of attention.
The simple understanding of this idea is that since everything can be tokenized, AI models can also be tokenized as an asset for issuance.
But making IMO work in practice is probably not that simple.
Quick understanding of the tokenization of AI models
For all ICOs and their variations, the key is to create a token, assign it various conditions such as quantity, release conditions, roles, and functions, and then form a market price.
The token here does not actually correspond to the real world, it can be created out of thin air, which is commonly known as "creating a coin."
But IMO is different.
The core point of IMO is actually the monetization of AI models in the real world.
Many open-source AI models face the challenge of monetizing their contributions, leading to a lack of motivation for contributors and organizations to earn money. This is also why the artificial intelligence industry today is mainly dominated by closed-source, for-profit companies. The key to the development of open-source AI models lies in raising more funds and building openly.
Therefore, the purpose of IMO is to provide a new way of issuing assets to help open-source AI models raise more funds to support their development.
Compared to some previous IXOs, if you are bullish on a tokenized asset and choose to invest in it, the market performance of the token may also give you a return, and the protocol corresponding to the token generates income, which you may also share; 
Now, in the IMO scenario, if you are bullish on a certain AI model, you can choose to invest in its corresponding token, and the provider of the AI model can raise funds for development and growth; at the same time, if the model generates economic benefits in actual use in the future, you may also share in the profits.
How does IMO work in practice?
To represent an AI model in token form and enable profit sharing, there are inevitably several key issues to address:
- How to ensure that a specific AI model is genuine and corresponds to the token you hold?
- How to ensure that the token holders can genuinely share in the profits generated by the AI model's use?
Ora Protocol uses two different ERC protocol standards, ERC-7641 and ERC-7007, and combines them with oracles and ZK technology to address the above issues.
- How to ensure that a specific AI model is genuine and not just a concept for fundraising in the token space?
First, it is important to note that Ora Protocol is an AI oracle protocol, and its core product is called the Onchain AI Oracle (OAO).
The role of this oracle is to verify and execute AI models on the blockchain, ensuring that the deployment and operation of AI models are fully conducted on-chain, thereby ensuring transparency and verifiability in their execution process.
However, because AI models are often a core competitive advantage, if they are exposed to everyone, they lose their competitive edge in business. Therefore, Ora Protocol also incorporates another technology - opML (Optimistic Machine Learning).
In simple terms, opML may use zero-knowledge proofs or other forms of cryptographic proofs to demonstrate that the model's operation results are correct without revealing the details of the model itself. This ensures the authenticity and effectiveness of the model while protecting its privacy and proprietary nature.
The specific implementation of opML is supported by the published paper in the above image, and we are unable to evaluate the technical details of its superiority or inferiority, but understanding the effect of this technology is sufficient.
Thus, through AI oracles and zero-knowledge proofs, we have addressed the issue of "how to prove the existence of a genuine AI model."
- The next issue is how to ensure that the ownership of the token corresponding to the AI model is yours and that you can share in the profits.
Tokenizing an AI model is the key to IMO. Ora Protocol introduces a token standard called ERC-7641, which is compatible with ERC-20.
If a developer of an AI model believes their model is good and wants to conduct an IMO in the cryptocurrency market, their approach is likely to be as follows:
First, associate the AI model with a specific ERC-7641 asset and specify the total quantity of tokens in the smart contract of that asset.
Second, investors in the cryptocurrency market purchase the token, and based on the quantity purchased, they correspond to the ownership proportion of the AI model (similar to being a shareholder).
Third, once the AI model runs on the chain, if the model or its content generates income (such as usage fees paid when the model is called, or royalties from NFT sales generated by AI), the ERC-7641 protocol can predefine the rules for distributing income in the contract and allow token holders to automatically distribute income based on their token ownership proportion.
Through this mechanism, ERC-7641 tokens become a bridge connecting the economic value generated by AI models and their token holders, allowing contributors and investors of open-source AI models to share in the long-term value of the model.
Therefore, ERC-7641 tokens are also known as Intrinsic RevShare Tokens, which can be interpreted as a token standard specifically designed for sharing the income generated by AI models.
So, the overall logic of IMO is very clear: AI model developers need to raise funds, associate the model with a token for IMO; buyers purchase the token and, according to the rules of the token smart contract, enjoy the sharing of the AI model's future use and creative work.
But at this point, there is still a critical loophole:
- How do you know that the AI works created on-chain later (such as NFTs, images, videos, etc.) genuinely come from the AI model that underwent IMO, and are not forged?
The solution provided by Ora Protocol is to mark these AI-generated works and implement it through ERC-7007.
Excluding technical details, ERC-7007 can be understood as a token standard designed specifically for AI-generated content to ensure the authenticity and traceability of the content's source.
This standard records the metadata of AI-generated content on the blockchain (such as the AI model used to generate the content, generation time, conditions, etc.) and uses smart contracts to automatically execute these verification logics. Developers can use zkML or opML to verify whether the AIGC data of a specific NFT truly comes from a particular machine learning model and specific input.
This increases the transparency of AIGC content authenticity and, through the immutability of the blockchain, ensures that once recorded, it cannot be altered or forged. Therefore, in the ORA protocol, ERC-7007 is also referred to as "Verifiable AI-Generated Content Token."
Currently, this standard is open source and can be found here.
Thus, we now fully understand the logic of IMO:
Bind AI models with tokens that have revenue-sharing functionality to conduct IMO
Investors enjoy a share of the future use and derivative creative work profits of AI models based on their token holdings
Use a token protocol for verifying content creation ownership to verify whether a work truly comes from a specific model and share profits
Still a Game of Assets, Not Perfect
From ICO to IMO, when AI models can also be tokenized for issuance, this year's cryptocurrency craze is bound to be closely tied to AI.
However, the IMO gameplay established by Ora Protocol is not without flaws.
Off-chain usage issues: Even if IMO can achieve the on-chain tokenization and profit sharing of AI models, it still struggles to address the issue of profit sharing when the model is used off-chain. When AI models are used in non-blockchain applications, tracking and distributing the profits from these uses to token holders is a complex problem.
Market demand uncertainty: While on-chain AI-generated content (such as NFTs) has brought new possibilities to the creative industry, the market demand for these works still carries significant uncertainty. The market value and liquidity of AIGC works, as well as how much people are willing to pay for these works, are unknown, making stable AI model profit sharing a distant prospect.
Actual effects of profit sharing: In theory, achieving profit sharing through ERC-7641 tokens sounds like an appealing idea. However, in practice, the effectiveness and feasibility of this mechanism still need to be tested by the market. Especially considering the high volatility of blockchain projects and tokens, the actual profits that token holders can obtain may vary significantly.
In the crypto world, people can play around with asset issuance, but whether the assets themselves are useful and how many people use them, rarely comes with a predetermined answer.
However, through this new mode of asset issuance, IMO does provide an innovative framework for open-source AI models to obtain funding support and achieve value sharing through tokenization.
This framework itself is a narrative closely tied to the hot topic and possesses positive value.
In a game of assets that is never perfect, riding the wave of AI's popularity often leads to success.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
