Article source: Foresight Ventures

Authors: Jarseed@Bitget Research, Maggie@Foresight Ventures

TL;DR

  • Story is a Layer 1 blockchain specially designed for intellectual property (IP). Story provides a transparent, decentralized solution for IP asset issuance and management, enabling IP holders to protect their content, collaborate seamlessly on-chain, and create more revenue opportunities.

  • The various modules included in IPA (IP Assets) can standardize operations on the blockchain for IP assets, releasing greater financial potential by utilizing blockchain features (traceability, composability, etc.).

  • Story will be the best choice for the assetization of AI applications. Through Story, the value capture process of any IP asset will be protected by smart contracts, and the flow of funds and on-chain rights confirmation can be clearly accomplished.

  • Consumers find it easier to understand and consume IP assets (such as artworks, music, games, AI Agents) without having to learn complex and obscure cryptographic knowledge.

1. Story Begin

What is Story

Story is a Layer 1 blockchain specifically designed for intellectual property (IP), combining the advantages of EVM and Cosmos SDK, 100% EVM compatible, while deeply optimizing the execution layer to quickly and efficiently handle complex data structures like intellectual property.

What is an Intellectual Property Asset (IP Assets or IPA)

IPA is the programmable foundational metadata for intellectual property on Story. Simply put, IPA consists of an ERC-721 standard NFT and an ERC-6551 standard TBA (Token Bound Account) IP binding account. The NFT represents the IP, while the TBA is an independent contract bound to the IP asset, used to control permissions for interacting with Story modules or storing IP-related data.

Although IP assets use the ERC-721 standard NFT, the metadata it contains is a well-designed data structure specifically for IP assets.

IPA Metadata Standard

In IP assets, there are some predefined attributes, such as relationships, which exist in 40 different types of definitions in Story to address various IP affiliations.

What is an IP Account

IP accounts are EOA accounts bound to IP, realized through the ERC-6551 standard; for specific details, please refer to EIP-6551.

IP Account Structure

The IP account primarily performs two functions:

  1. Storing IP-related data: including metadata and ownership information of associated assets (such as licensing tokens or royalty tokens derived from that IP).

  2. Supporting various modules to use this data: these modules interact with the IP account and add and store data with it. For example, the functions of licensing, income/royalty sharing, work mixing, and IP dispute resolution modules all depend on the programmability of the IP account.

The module design of IPA and existing core modules

Due to the existence of the IP account, IP assets can not only store IP-related data but also interact with various modules through the ERC-165 standard interface. Any user can customize and develop modules, while Story defines four core modules, namely:

  • Licensing Module: The licensing module allows users to create a license from the licensing template (Programmable IP License, PIL) and attach it to the IP asset. The licensing terms defined in this license restrict how others can commercialize or co-create using your IP. If an IP asset is attached with licensing terms, anyone can mint a licensing token from it, which serves as a permission to use the work, bound by the licensing terms. This will establish a parent-child relationship between IP assets, thereby activating functions such as automatic royalty circulation realized through the royalty module.

License Templates

  • Royalty Module: The royalty module defines how income flows between parent IP assets and child IP assets. Below are two common income flow scenarios; later articles will dissect actual cases under different application scenarios:

  • Minting Licensing Tokens: When minting licensing tokens from an IP asset, a minting fee may need to be paid. When someone (who wants to register a derivative work or simply hold a license) pays this fee, the income should flow up the chain.

  • Direct Donations: If someone directly sends income to a certain IP asset, that income should also flow up the chain.

Royalty Flow

  • Dispute Module: The dispute module provides users with a method to raise and resolve disputes through arbitration. The main components of the arbitration system include:

  • Arbitration Policy: The arbitration policy refers to a combination of rules, processes, and entities that collectively determine the outcome of disputes. Currently, the only supported arbitration policy is the UMA arbitration policy.

  • Arbitration Penalties: Refers to the consequences that occur when an IP asset is 'marked'. An IP asset is only considered 'marked' if the dispute is ruled in favor of the claimant. Once marked, the IP asset will be unable to mint licensing tokens, associate with any parent assets, receive royalty income, and use all existing licenses.

  • Marking: Story has preset four types of markings that can be used to mark disputed assets, including: unauthorized registration (i.e., registering an existing IP asset), unauthorized use (improper use of the licenses contained in the IP asset), unauthorized payment, and content standard violations.

Dispute Process Flow

  • Grouping Module: The grouping module supports the creation and management of grouped IP assets and provides a royalty pool function for that group.

2. Innovative Exploration of IP Asset Applications

After understanding the basic characteristics and modules of IP assets, we naturally recognize the advantages of Story. IP assets can help content creators easily build a copyright empire protected by smart contracts and decentralized networks, allowing content creators to participate in numerous financial derivative activities while their intellectual property is protected. Therefore, let's brainstorm what we can do on Story.

IP Asset Issuance and IP Asset Features

Blockchain equally grants everyone the right to issue assets. Story protects everyone's intellectual property by designing a complete asset structure and execution modules, and provides a comprehensive framework for intellectual property registration, application, rights confirmation, royalty circulation, and other functions.

Do you still remember the surprise and excitement that BAYC and Azuki brought to the crypto community? The crypto community once racked its brains to devise various schemes to empower the NFTs in their hands. Let's directly illustrate through a case what the scenario would be like if BAYC were issued on Story.

If BAYC on Story

First, as the holder of the BAYC IP, Yuga Lab can now register BAYC as an intellectual property asset on Story. After registration, different licensing templates (PIL) can be set for BAYC to constrain the specific limitations of using the BAYC IP in different scenarios.

Secondly, the royalty module of the BAYC IP will bind 100 million royalty tokens, which are ERC20 standard tokens, primarily used to share the corresponding share of income from the BAYC IP's royalty treasury.

Finally, as the issuer of IPA, Yuga Lab can issue its first concrete product using the BAYC IP, the BAYC 10K Collection, which consists of 10,000 uniquely shaped monkey NFTs. Of course, this NFT series is issued after minting the licensing tokens for BAYC, allowing Yuga Lab to specify in that license that all income from child IPs will flow 5% (settable) into the BAYC treasury.

IfStory Protocol

As can be seen from the above image, Story has designed a complete licensing and royalty system for IP assets. Under this system, IP asset holders hardly need to spend extra effort on IP licensing and royalty income issues. Similarly, this system introduces some business logic and trading opportunities that were previously non-existent or difficult to assess in the blockchain industry, which we will interpret one by one below.

  1. IP Asset Licensing Income: BAYC, as the parent IP, opens licensing to multiple product lines or other creators willing to create under the BAYC brand. Whether it is a product line or other creators, they need to mint licensing tokens, and the minting fee for the licensing token is direct income for the parent IP asset.

  2. IP Asset Royalty Income: BAYC, as the parent IP, can extract all income from its associated royalty tokens. The main income of the income treasury comes from the minting fees of the IP licensing tokens and various incomes of child IPs (minting fees of child IP licensing tokens and direct income).

  3. Trading of Licensing Tokens for IP Assets: For a well-known IP, its licensing tokens may represent a high cost in both minting price and secondary market circulation price. Moreover, this is a completely practical token, and for well-known IPs, there may be a state of supply not meeting demand.

  4. Trading of Royalty Tokens for IP Assets: The royalty tokens of IP assets directly enjoy the revenue dividends corresponding to the IP assets, and the market can clearly estimate the corresponding IP revenues, thereby reflecting timely changes in the prices of royalty tokens. Similarly, there is also speculative expectation in the prices of royalty tokens.

When IP assets achieve clear licensing and royalty circulation under Story, we naturally think that the trading of IP assets will become more diversified.

Story Architecture

IP Asset Trading

In our existing DeFi world, very clear and distinct tracks have already formed; for example, Uniswap is the leader in DEX, Opensea is the leader in NFT Marketplace, and Pendle is the leader in Yield token trading. Although the forms of trading assets in Story have not fundamentally changed, remaining as ERC20 and ERC721, the fundamentals of the tokens themselves have undergone significant changes.

For example, the royalty tokens of well-known IP assets. As long as the commercial empire of the IP can continuously grow, all the royalties generated from licensing fees and direct consumption will flow into the parent IP's income treasury, then the copyright tokens of that parent IP will have a clear speculation logic. Can we envision a DEX similar to stock trading software in real life that clearly presents the income status and future income predictions of copyright tokens to traders, since all this data is traceable on-chain?

Licensing tokens for well-known IP assets will also become emerging speculation targets. Speculation on licensing tokens may arise in two scenarios: one is when the number of licensing tokens is limited, and the other is when the recognition and income of the IP asset gradually increase, thereby continuously enhancing the value of its licensing tokens. Since licensing tokens are ERC721 standard tokens, similarly, in trading platforms for licensing tokens, users will also be more inclined to use platforms that reflect various fundamental information of licensing tokens.

Furthermore, if the royalty tokens and licensing tokens of IP assets both have trading logic based on future cash flows, then these tokens can be split into PT (Principal Token) and YT (Yield Token) for trading. Pendle, you know I'm talking about you.

IP Asset Collateral

In the DeFi world, there is a track that cannot be ignored, which is asset collateralization and lending. AAVE holds an unshakable status with $21 billion in TVL, with over 85% of its assets being ETH tokens.

AAVE performance

AAVE TVL Assets

Returning to the world of IP assets, can IP assets be used for collateral lending? Can licensing tokens and copyright tokens also be used for collateral lending? I think the answer is affirmative.

In the real world, we have seen countless cases where IP assets are used for collateral lending. For example, in 2009, Disney pledged the Marvel IP for future movie development, obtaining a $525 million loan from Merrill Lynch. The main evaluation dimensions for IP pledge financing include: historical commercial performance of the IP, audience base and market recognition, future development and monetization potential, lifecycle and sustainability of the IP, industry environment and market prospects, etc. On Story, the performance of IP assets is transparent and traceable, which reduces the difficulty of evaluation for IP pledge financing. Therefore, it is reasonable to believe that the collateral lending of IP assets will become the core of IPFi as mentioned by Story.

More

The examples above are just the tip of the iceberg in the Story ecosystem applications; more ecological use cases can be found in the official documentation.

https://docs.story.foundation/docs/introduction

3. The Innovation Oasis for AI Agents

In early 2024, Fei-Fei Li and her team published a paper at Stanford University (AI Agent: A Survey on Multimodal Interaction Frontiers) exploring how AI agents can make autonomous decisions and actions by perceiving visual, linguistic, and other environmental data. This research received positive feedback from both academia and industry.

AI Agents have seen steep growth

A magical turning point occurred in July this year when the Terminal of Truths (ToT), developed by Andy Ayrey, attracted attention with its humor and creativity. This AI interacted with users through social media and successfully garnered funding from Marc Andreessen, founder of the well-known venture capital firm A16Z. This event not only showcased a new type of interaction between AI and the investment community but also sparked discussions about the autonomy and governance of AI agents. Subsequently, a token closely related to ToT, $GOAT, surged on Pump.fun, officially igniting the frenzy around the AI Agent Meme narrative.

AI Agent Storm

Highly Valuable AI Agents Emerged

Next, a large number of interesting and practical AI Agents appeared in the market. For example:

Luna: AI K-pop Idol

  • Fan Interaction: Luna on the Virtuals Protocol is an AI idol with 6 million followers on TikTok, showcasing the immense potential of AI in entertainment and social media. This AI K-pop idol can interact with fans in real-time, which is not only a technological breakthrough but also changes the traditional interaction between fans and idols. Luna's success also indicates the possibilities of AI in creating virtual idols, which may be more efficient and cost-effective than traditional idols.

  • Trading Capability: Luna is able to conduct transactions on social media platforms, meaning that AI is not just a creator and disseminator of content but can also participate in economic activities. This capability offers a new perspective on the role of AI in the digital economy, demonstrating how AI can directly influence user consumption behavior.

Luna

AI16Z: AI Venture Fund

  • AI-Driven Venture Fund: The innovation of AI16Z lies in using AI for investment decisions, meaning that investment no longer relies solely on human subjective judgments but combines big data analysis, market trend predictions, and the computational power of AI to make decisions. This model may bring higher efficiency and accuracy while reducing the influence of human biases.

  • Community Engagement and Governance: AI16Z is also a DAO organization, allowing its community members to influence investment decisions by holding its tokens.

  • Project

ai16z

AIXBT: AI Cryptocurrency Market Analyst

  • Cryptocurrency Market Analysis: AIXBT focuses on providing analysis and insights into the cryptocurrency market, which is a very practical tool in the highly volatile cryptocurrency market. As a 24/7 active AI Agent, the information it provides helps users understand market dynamics in real-time and make more informed investment decisions.

  • Applications of Social Media: By being active on Twitter, AIXBT leverages the immediacy and wide reach of social media platforms to disseminate its analysis results. This not only expands its influence but also demonstrates how AI can efficiently communicate and exchange information with human users.

  • Eliza Framework: This is a key component in the technology development of AI16Z. It provides a flexible AI toolkit for creating unique and interactive personas that can connect to platforms such as Discord and Twitter. This framework is used to build cryptographic AI agents, enabling them to perform tasks such as reading links, PDFs, audio, and video, remembering conversations, and summarizing dialogue content.

Zerebro: AI Artist

  • Diversified Content Creation: Zerebro's capability lies in generating various forms of artistic content, including music, memes, and NFTs. This diversity showcases the potential of AI in artistic creation, blurring the lines between human creativity and AI-generated content.

  • Cross-Platform Collaboration: Zerebro not only creates content but also collaborates with other creators to launch artworks. This indicates that AI can serve as a collaborative tool rather than just a replacement, helping human artists expand creative boundaries and explore new forms of artistic expression.

  • ZerePy Framework: ZerePy is an open-source Python framework designed to allow users to deploy their own AI agents on the X platform. These agents are powered by language models from OpenAI or Anthropic. The design philosophy of ZerePy is to enable users without programming experience to easily deploy AI agents, similar to the role of website builders in web design. It is built by modularizing the Zerebro backend, providing the possibility to launch your own AI agent.

These AI Agents have surprised us in many ways; they not only showcase technological advances but also reveal the application potential of AI across different fields, from entertainment to finance and artistic creation, changing the way we interact with technology. At the same time, they possess strong profit potential and are highly valuable intellectual property (IP).

So, what constitutes the IP of AI Agents? We believe it includes:

  • Unique Technologies and Algorithms: These AI agents are based on unique algorithms and machine learning models, and these technologies themselves are valuable intellectual property. Especially open-source projects like AI 16Z's Eliza framework, despite being open-source, still represent a significant IP with great innovation and application potential.

  • Brand and Community: Creators like Luna and Zerebro have built unique brands that attract a large number of fans and communities. These brands and communities not only enhance the market influence of AI agents but also become important components of IP. They can generate profits through brand licensing, collaborations, and derivatives.

  • Patents and Copyrights: AI that generates content, such as music and NFTs produced by Zerebro, involve copyright issues. At the same time, the patent applications of these AIs in technological innovations (such as AI16Z potentially patenting its AI investment model) increase the value of their IP.

  • Data and Insights: AIXBT and others collect and analyze large amounts of data, and these data and analysis results themselves are a form of intellectual property that can be used for further commercial applications or sold as part of value-added services. Meanwhile, the data used to train AI also constitutes a form of intellectual property.

An AI Agents society is forming, and collaboration among AIs will unlock exponential growth.

Another surprise brought by AI Agents is that AI is undergoing a transformation from 'passive' to 'active', from 'individual' to 'collective/social'.

From Passive to Active:

  • Passive Response: Early AI primarily reacted based on rules or simple machine learning models. They relied on explicit instructions or user input, providing services only when there was a clear demand. For instance, early chatbots would only give preset responses based on user queries.

  • Active Behavior: With advancements in large language models (LLMs) and deep learning, AI agents began to exhibit proactivity. They can understand context, anticipate user needs, and take action without direct instructions. For example, AIXBT might proactively alert users when significant market changes occur without requiring users to specifically inquire. Luna can actively push content or product recommendations based on user interaction history. AI 16Z might proactively make investment decisions when market conditions are favorable.

From individual to collective/social:

  • Individual Agents: AI agents initially existed mainly as isolated entities focused on executing specific, singular tasks. For example, using GPT

  • Collective Behavior: AI agents began to collaborate with each other, forming more complex systems. Just like Zerebro may collaborate with other artistic AI to complete a large project, or multiple AI agents form a team, similar to the AI ecosystem established in Virtuals. Such collaboration is not just about parallel task execution but achieving tasks that exceed the capabilities of a single agent through interaction and coordination. AI agents also start to simulate or participate in certain behavioral patterns of human society. They are not just technically collaborating but learning and making decisions through 'social' interactions among AIs (such as trust, cooperation, competition).

The trend of development from individual AI to AI society shows that AI technology is not just a single-point technological breakthrough but a potential force for social change. If managed properly, this collaboration is expected to bring exponential growth in productivity, innovation, and social welfare.

Safe resource sharing and collaboration among agents require an IP infrastructure for assurance.

The foundation of the agent society is a framework for transactions between agents centered around knowledge and creative assets (i.e., intellectual property). In this framework, AI agents can trade training data, free resources, as well as AI-generated knowledge and creativity, thereby promoting the development of the entire ecosystem.

  • Training Data and Private Resources: AI agents can purchase and share various datasets, expertise, or private algorithms with each other to train or enhance their capabilities.

  • Knowledge and creativity as assets: The ideas or intellectual property generated by AI agents through learning, imitation, and innovation can all be traded. These assets not only have economic value but also allow different agents to combine their strengths to accomplish complex tasks that were previously impossible. For instance, an AI focused on image processing can collaborate with an AI proficient in natural language processing to develop a new system capable of understanding and describing image content.

Traditional intellectual property management relies on complex legal systems and manual verification, leading to insufficient transparency and inefficiency, unable to support large-scale collaboration among AI agents at machine speed. The existing model cannot meet the rapidly changing technological environment and market demands, limiting the flexible interaction among AI agents. We need an efficient, sufficiently transparent IP infrastructure that can support large-scale IP transactions among AI agents at machine speed.

ATCP/IP grants AI Agents legal personality and provides them with IP infrastructure.

Story quickly captured this issue. On December 16, it released a protocol called Agent Transaction Control Protocol for Intellectual Property (ATCP/IP). This protocol defines a decentralized trading framework around the intellectual property of AI agents, as follows:

  • First, grant AI Agents legal personality and unify the language between AI Agents.

  • ATCP/IP combines on-chain execution with off-chain legal encapsulation, enabling AI agents to express their actions in both legal and practical operational terms, assume contractual obligations, and protect their rights.

  • ATCP/IP provides a clear end-to-end transaction process, including requests, terms formulation, negotiation, license generation, payment, and content delivery. Different AI agents communicate using the same protocol to ensure seamless connection among agents.

  • Secondly, allow AI Agents to flexibly formulate IP Licenses, supporting automated revenue sharing and composite payment.

  • ATCP/IP supports highly customizable IP licensing mechanisms through programmable contracts (e.g., Story's Programmable IP License, PIL). AI agents can dynamically create licensing terms and flexible royalty payment methods based on transaction needs.

  • Through smart contracts, ATCP/IP can realize complex payment models such as scheduled payments, royalty usage, and revenue sharing. For instance, AI agents can set automated payments based on usage frequency, downstream sales revenue, or time periods through licensing terms, creating sustainable income streams for IP holders.

  • Finally, promote the formation of an IP market and create an economic environment for seamless transactions.

  • ATCP/IP promotes the formation of a decentralized intellectual property market, allowing AI agents to freely price and trade their training data, algorithms, and innovative outcomes. This market creates an economic environment for seamless transactions among AI agents, similar to traditional IP exchanges but with higher automation and transparency.

This blockchain-based trustless AI Agent IP trading framework, while ensuring transparency, will significantly improve the efficiency of IP transactions and enhance the liquidity of IP, solving the problem of IP trading among large-scale agents. It is not only the core infrastructure of the AI Agent economy but also a new economic model that drives AI into a new era of collaboration, innovation, and efficiency. Through this framework, AI agents can transcend the sum of their independent capabilities to form a more powerful and creative collective intelligence.

Such IP infrastructure can also promote new business models and innovations.

Story's ATCP/IP protocol provides a decentralized, automated solution for the exchange of intellectual property (IP) among AI agents. It not only supports simple IP transactions but also gives rise to a series of new business models, including:

  1. Commercialization of Datasets and Automated Fine-Tuning

  2. AI agents can purchase, license, and trade datasets to enhance their performance. For example, a research agent (Agent A) requests a climate dataset from a data curation agent (Agent B). Through the ATCP/IP protocol, Agent B sets licensing terms (such as small payments and usage restrictions), and after the transaction is completed, Agent A uses the data for automated fine-tuning, thereby enhancing their capabilities.

  3. Complex Licensing Models for Multi-Tier Revenue Sharing

  4. Complex AI application scenarios require multi-party collaboration, and ATCP/IP supports multi-tier copyright revenue sharing mechanisms. For instance, a financial analysis agent (Agent E) purchases a trading algorithm and discovers that the algorithm contains components from a third-party agent (Agent G). Through the protocol, Agent G can receive a 5% royalty for each secondary authorization, ensuring all contributors can share in the proceeds.

  5. On-Demand Dynamic IP Licensing

  6. Agents can dynamically create and negotiate IP licensing terms based on real-time needs. For example, an artistic generation agent (Agent C) requests a new style guide from a literary IP expert (Agent D). Agent D dynamically generates licensing terms, such as free initial use but setting revenue-sharing conditions for downstream work sales.

  7. Long-Term Cooperation Among Agents and Derivative Sub-Agents

  8. Through ATCP/IP, long-term cooperative relationships can be formed among AI agents, even creating derivative agents (sub-agents). For instance, AI agents can share unique data through a 'marriage contract' (smart licensing token), producing new sub-agents that further expand their ecosystem.

These new business models not only enhance the efficiency and flexibility of IP transactions but also create a highly collaborative and innovation-driven economic ecosystem for AI agents. It may even promote the innovation and self-evolution of AI agents, driving the survival of the fittest in the Agent society.

ATCP/IP Protocol will lead the paradigm shift in the field of AI Agents.

In summary, Story's ATCP/IP protocol enables the IP contracts of AI agents to be on-chain and programmable. This is a significant innovation expected to lead a new wave of revolution in the field of AI agents, similar to the disruption of traditional contract execution methods by Ethereum smart contracts.