Author: CloudY, Jam
Editor: Vincero, YL
Review: Yasmine
At the end of November 2022, OpenAI launched the intelligent dialogue system ChatGPT, and this news quickly attracted global attention and heated discussions.
AI-related stocks, whether A-shares, US stocks or Crypto, have all seen a surge. In the following years, as ChatGPT was widely used, people realized its impact on the entire world, and new ChatGPT application scenarios and similar product iterations continued to emerge.
Even Microsoft has gained unlimited expectations from investors for acquiring OpenAI and embedding ChatGPT into Bing and Office ecosystems, and its stock price has soared. Until the emergence of ChatGPT4, the super artificial intelligence it demonstrated made people calm down from the excitement and start thinking about the changes that AI will bring to their own industries, as well as the potential risks after further development of AI.
This article is written in this context, and tries to find the answer to this question through research and thinking on the two industries of AI and Blockchain.
The current status of AI industry development
Productivity Tools
AI can be seen as a productivity tool. Just like stone tools, steam engines, internal combustion engines, electric motors, computers and the Internet to humans, they can be used as tools to bring about dramatic changes in productivity and production relations in human society. AI changes productivity. It lowers the threshold for computer interaction for humans and also improves the efficiency of repetitive production for humans. The former improves the quality of human life, and the latter reduces obstacles to human development.
Specifically, AI technology has had a profound impact in many industries, such as smart manufacturing, healthcare, finance, transportation, education, etc. Through AI technology, people can let machines learn and perform some non-creative tasks autonomously, which can improve production efficiency and reduce costs for some industries. For example, using AI to predict protein structure to develop new drugs, ESMFold trained by the Meta AI team has predicted more than 600 million metagenome proteins, showing the breadth and diversity of proteins in nature, which was unimaginable in the past.
To put it simply, using AI technology, we can use natural language to process complex programs in the past. We do not need to understand the principles of complex programs, nor do we need to know how to write code. We only need to tell AI what kind of result we want, and AI can execute the intermediate steps based on this result to achieve a desired result. This is the improvement in productivity brought about by AI.
This is the well-known AIGC, which will be more widely used in the fields of intelligent customer service, virtual humans, games, etc. ChatGPT can provide a smoother and more natural conversation experience for virtual humans, games and other fields based on the existing corpus, further improving the user experience and the market competitiveness of the product. More importantly, ChatGPT can replace humans to complete some repetitive content creation, such as programmed reports, simple information collection and summary, translation, and illustration drawing with limited conditions. Further liberate human productivity and focus on entering key instructions or creating, rather than repetitively executing instructions.
Technology Trends Guide
The current core applications of AI include general artificial intelligence, knowledge graphs, data analysis and synthesis, autonomous driving, and AIGC.
in:
Knowledge graph: Knowledge graph represents various entities, relationships and attributes in the form of a graph to support applications such as intelligent search, recommendation and question-answering.
Synthetic data: Synthetic data is data generated by machine learning and other AI techniques, which can be used to train and evaluate AI models. Due to privacy and security reasons, real data is often difficult to obtain or share. Therefore, synthetic data can replace real data in some scenarios.
AIGC: AIGC technology is a technology based on deep learning and generative models. It can be used in multiple fields such as text generation, audio generation, image generation, video generation, etc. It is also the most widely discussed and applied direction at present.
Whether in terms of the number and amount of market financing or media attention, 2022 is undoubtedly the year of AIGC explosion. However, AIGC is still a relatively new technology and is still in the early stages of exploration and development.
Specifically, the development stages of AIGC can be described as:
Research phase: Focus on the basic principles and algorithms of AIGC, explore how to train and optimize models, and establish databases.
Application stage: AIGC begins to be applied to various practical scenarios and begins to explore how to apply AIGC technology to specific fields.
Industrialization stage: AIGC begins to be widely used in various industries and fields, forming its own industrial chain and supporting ecosystem.
Overall, we have just moved from the research stage to the application stage, and the development of AIGC is still in its infancy.
the core element
Data, algorithms and computing power are the three core elements in the development of AI.
In terms of data, with the continuous development of AI technology, the quality and diversity of data are becoming more and more important. In addition to massive application scenario data, data also needs to be effectively cleaned, preprocessed and labeled to improve the training accuracy of the algorithm. In addition, cross-modal and cross-domain data fusion issues need to be considered to better explore the value and intelligence of data.
In terms of algorithms, the development of AI technology is still in the process of continuous iteration and improvement. Future development trends will mainly be reflected in the multimodal and large models of deep learning algorithms, as well as innovations in autonomous learning, knowledge transfer, and incremental learning. This will further improve the intelligence level and application scope of AI algorithms and promote the widespread application of AI technology.
In terms of computing power, as AI computing continues to accelerate and optimize, hardware carriers are also constantly upgrading and improving. For example, the emergence of dedicated chips such as GPU and TPU has greatly improved the efficiency and speed of AI computing. In addition, the development of cloud computing and edge computing has also provided a more flexible and diverse computing environment for AI computing power.
The current status of the Blockchain industry
Distributed Ledger
Blockchain is a decentralized distributed ledger.
First of all, Blockchain has the property of being tamper-proof, which comes from the consensus mechanism at the bottom of the blockchain. Since the on-chain data is recorded by blocks and witnessed by miners/verification nodes, and blocks are connected and recorded continuously, the on-chain data generated by smart contracts and accounts cannot be modified once recorded in blocks.
As the number of nodes increases, geographical locations become more dispersed, computing power increases, or the value of staked tokens increases, it becomes more difficult and costly to disrupt consensus. Therefore, it is difficult for a centralized individual to change what has been recorded.
Secondly, under the premise of being tamper-proof, smart contracts built on code allow users to interact with them without having to trust anyone. Smart contracts will run the code according to the preset path to implement corresponding operations. This makes trustless on-chain transactions possible.
At the same time, only the corresponding account can call the assets in the smart contract, and there is no situation where other accounts transfer the assets of the original account through the smart contract. Because every operation of the original account requires a signature to confirm the identity, and the first transfer interaction even requires Approve the smart contract to call the account assets. This makes the user's wallet account the best carrier of its identity (DID) and assets.
Within the framework of consensus mechanism and smart contract, all on-chain assets and on-chain behaviors can be recorded and confirmed, and the rights and interests generated based on them can be automatically collected into the accounts of their owners. This can directly solve the problems of "real and fake Monkey King" and "Li Dai Tao Jiang". No one can steal other people's assets by simply copying and pasting, and no one can replace the rights and interests of the owner to obtain their interests.
Specifically, digital assets can be defined in the form of tokens with their unique smart contract addresses, such as using NFTs to represent digital paintings; and anyone's behavior can be proved by non-tradable tokens (SBT), such as authenticating their work content or space-time existence (Proof of Work/Proof of Attendance).
Technology Trends Guide
Layer 0-2 is the hierarchical structure of the Blockchain technology architecture, while consortium chain and private chain are different types of Blockchain application scenarios.
Layer 0: Layer 0 refers to the physical facilities and network architecture of Blockchain, including hardware equipment, network protocols and transmission media, etc. It carries the underlying role of cross-chain information and asset resolution. Currently, Cosmos, Polkadot and LayerZero are the main technical representatives.
Layer 1: Layer 1 is the basic layer of Blockchain, also known as the public chain, including Bitcoin, Ethereum, etc. The protocol design and technical implementation of Layer 1 determine the basic performance and functions of Blockchain. According to the type, it can be divided into EVM and non-EVM systems.
Layer 2: Layer 2 refers to the protocols and solutions built on top of Layer 1, which are used to improve the performance of Blockchain and expand application scenarios. There are currently 6 technologies for Layer 2 protocols, among which ZK Rollup and Optimistic Rollup are the mainstream. These protocols can enable Blockchain to process more transactions, increase TPS and reduce Gas fee, etc.
Consortium chain: A consortium chain is a blockchain network jointly managed and controlled by multiple organizations or institutions. These organizations usually cooperate for common interests, such as banks, insurance companies, supply chain companies, etc. Consortium chains are different from public chains in that they have limited participants and a relatively small number of nodes, so their transaction speed and security have been improved to a certain extent.
Private chain: A private chain is a blockchain network independently controlled by a single organization or institution, usually only allowing internal personnel to participate.
the core element
Distributed nodes, cryptography, consensus algorithms, smart contracts, and cryptocurrencies are the core elements of Blockchain development.
Distributed nodes are the core of Blockchain technology, which enables data to be stored and transmitted in a decentralized manner. Cryptography is an important theoretical tool to ensure the security and privacy of Blockchain. In addition, consensus algorithm is the key to achieve distributed consistency of Blockchain. Smart contract is a computer program that can be executed automatically and can execute various logical instructions on Blockchain. Finally, cryptocurrency, that is, the security and anonymity of transactions are guaranteed by using encryption technology.
Through distributed nodes, all participants can have a complete copy of the data, thus ensuring the transparency and security of the data. The core technologies of Blockchain - hash functions, digital signatures and asymmetric encryption are all applications of cryptography. They can help ensure the integrity of data and identity authentication, while also protecting the privacy of users.
Through the consensus algorithm, all nodes can reach a consensus, ensuring the consistency and immutability of data. Common consensus algorithms include PoW, PoS, etc. Smart contracts can realize transactions without the need for third-party trust, thereby improving the efficiency and security of transactions to a certain extent. The emergence of cryptocurrencies such as Bitcoin and Ethereum has promoted the widespread application and development of blockchain technology.
The intersection of Blockchain and AI
As part of the Blockchain industry, under the wave of AI, we also need to think: Does AI’s change to the world include Blockchain? If so, what will the change be? And what impact will Blockchain’s decentralization and rights confirmation capabilities have on AI?
First of all, AI as a productivity tool can lower the technical threshold, so it can naturally lower the technical threshold in the Blockchain industry and increase its production efficiency.
Secondly, AIGC will also free games and the metaverse from programmed settings, bringing new narratives and gameplay to Blockchain.
Blockchain's smart contracts will be able to define the areas and scope that AI can involve, or limit the authority of AI to prevent its overdevelopment.
At the same time, the decentralization of Blockchain can provide AI with resource sharing and allocation of underlying data and computing power required for training models.
In addition, Blockchain's property rights confirmation capabilities can also provide proof of data, identity, and ownership, avoiding conflicts of interest caused by AI.
What does AI mean to Blockchain?
First, AI as a tool can lower the threshold for content creation. It allows every ordinary person to show their creativity without technical restrictions and output high-quality content or NFT works. This includes but is not limited to NFT creation, game asset creation, metaverse modeling, code construction, etc.
However, the current application of AIGC in the NFT field is only simple image output, which is not fundamentally different from traditional Generative Art. The real application of AIGC in NFT should be a further expansion of NFT characteristics, just like Mirror World uses AI to build the soul of NFT.
Secondly, it is to lower the technical threshold for code writing. The code is divided into two directions, one is to issue projects and deploy smart contracts, and the other is hackers or white hackers. These two directions belong to the two ends of adversarial generation, that is, we can use AI for natural language programming and deploy the smart contracts we need, and the other party can also use AI to analyze the contract code and launch attacks. In this way, we will be able to use AI to iterate the deployed contract code, thereby forming an involution and helping the entire industry to build a more complete and reliable code. On this basis, everyone can put more thoughts on optimizing the architecture of the blockchain or designing the entire project, or on the economic model, to enrich the gameplay of the project and innovate the entire business level.
Similarly, when AI simplifies the entire technical threshold, the complex operations in the past will be widely used. For example, revolving loans, flash loans, optimal mining strategies, automatic income acquisition, and the judgment of the time to leave the first mine, the entire path can be completed by AI. AI can program itself, choose the path, and deploy it directly. Just like the skill card of Yu-Gi-Oh, we only need to use the skill card, and then the skill will appear and produce effects. This can delegate operations that required a high threshold to complete in the past to ordinary users. Take MEV as an example. If we want to obtain the value of MEV, we need to write a MEV clamp robot. When ordinary people can do it, there is no profit space, because when everyone can do it, they need to compete for Gas to get ahead. Due to the principle of game theory, the value of MEV will be squeezed out by the high Gas fee in the end, which will eventually lead to unprofitability and reduce the impact of MEV. This is a kind of technology decentralization that forces industry optimization.
Or it could promote the popularization of blockchain technology. According to Footprint Analytics data, there are only 320,000 active users of Ethereum, which is less than a fraction of Internet users. The biggest problem is that users do not have the need to enter the blockchain, and the few users who have the need are blocked by the complex on-chain interactions. In addition, in the past, data was uploaded to the blockchain or tickets and certificates based on the blockchain required the establishment of a blockchain system or the payment of a large amount of gas fees, which was a huge cost. Now, based on AI technology, we can realize blockchain construction at a low cost or optimize the on-chain data usage path to reduce gas fees. Therefore, anywhere that requires confirmation of rights and information transparency, blockchain technology can be used and smart contracts can be deployed. Therefore, the simplified interactive system through AI will bring a large number of users to the Blockchain industry.
What we need to know is that the changes that AI can bring only exist in the application layer of the blockchain. Based on their own cognition in the interaction, users use AI to skip the process of writing smart contracts and directly deploy applications to solve a certain need. The key to issuing projects will no longer be issuance, but innovation and operation. I believe that the pattern of the application layer will inevitably undergo earth-shaking changes in the future. However, AI is powerless to change the execution layer, consensus layer, and data layer under the application layer, because this is an innovation of the underlying mechanism, and it is definitely not a field where simplifying repetitive work can bring about qualitative changes. Just as the implementation of EIP1559 in the London upgrade gave Ethereum further impetus to move forward, the completion of the Shanghai upgrade can increase the ETH pledge amount, improve Ethereum security, and let the LSD sector take off again.
The role of blockchain in AI
The decentralized nature of Blockchain deviates to a certain extent from the centralized nature of current AI technology development, but it is precisely this deviation that provides a solution to the problems facing AI.
Modern AI and big data technologies are largely centralized, that is, they are usually controlled by a few large companies or organizations that have powerful technology and resources and the power to determine market trends and user behavior. This centralized nature requires people to trust that AI will honestly follow instructions when using it. Therefore, there are certain risks and problems in the development and application of AI, such as privacy leaks, algorithmic bias, data abuse, etc.
However, the distributed and decentralized nature of Blockchain can solve these problems. Through smart contracts, the data sets that AI can use and the scope of its operation can be limited to prevent AI from doing evil. At the same time, nodes can be established to monitor the behavior of AI. If it does evil, the supervisor can report it and confiscate the computing power used by AI, so that AI only does things that promote human development and prevents excessive use and unauthorized behavior of AI.
Specifically, for the sharing and confirmation of the underlying data required for AI training, Blockchain allows users to choose whether to provide their own data for AI model training. This will require the further development of zk technology to provide user data without disclosing personal information. The entire process of data collection, storage and sharing will be built on decentralized nodes to ensure data security and availability, and to confirm the source of data. As a result, when the AI trained using these data generates income, part of the income can be distributed as dividends to the data owner based on the data involved, ensuring the rights and interests of the data provider. The generation and sharing of AI training data mentioned above can also rely on the decentralization, security and transparency of Blockchain to ensure data privacy and security.
Users who provide prompts for AI operations can also obtain part of the revenue generated by the invocation of their prompts based on their ownership of the prompts, thereby ensuring the interests of both the AI data owner and the AI operation prompt owner.
Another thing worth mentioning is computing power mining. In addition to a large amount of data, the establishment of AI models also requires computing power for training, but the computing power in the world is currently in short supply. Then the computing power can be centralized in a decentralized form to establish a cloud computing mining pool, subsidize computing power providers through computing power mining, and then sell it to AI model training in the form of auctions, so that limited computing power can be maximized. At the same time, the security and reliability of computing can be guaranteed. More importantly, we can also integrate data, algorithms, and computing power to establish an AI as a Service protocol, and provide AI model building services to users in need with its own decentralized advantages and reusability. Then from data acquisition to data processing, to algorithm selection and computing power call, all are carried out through an ecosystem, which can ensure the advantages of the supply chain while avoiding centralization risks.
In addition to the construction of AI models, when we look at the application of AI, we can find that the problems of piracy, plagiarism, virtual people, etc. brought about by AI's super learning ability are not a problem in front of Blockchain. Artworks are recorded on the chain in the form of NFTs, and their unique smart contract address can prove the authenticity of the work. The value of an artwork depends not only on the artwork itself, but also on the identity of its creator. Just like the imitation of Van Gogh's sunflowers by later generations is worthless, but the blockchain can prove which sunflower was created by Van Gogh. The knowledge graph, one of AI applications, can also build a distributed knowledge graph based on the blockchain and ensure that the data in it will not be tampered with, deleted or misappropriated.
OAT or SBT can also be used to solve the problem of AI using personal past data to build virtual personalities. Any on-chain behavior has a corresponding record, and the OAT or SBT created for the relevant record is also unique. The identity can be defined based on the OAT or SBT in the account. This is all due to the immutability of the blockchain. What happened in the past is recorded in every subsequent block, and it is impossible to fabricate behaviors that did not exist in the past.
In summary, AI as a productivity tool can accelerate the development and popularization of the Blockchain industry, and AIGC has brought new directions and narratives to the Metaverse and NFT sectors, but AI can only replace repetitive work and lower technical barriers, and cannot innovate key technologies. Therefore, the changes that AI brings to Blockchain will only remain at the application layer.
Blockchain is a risk controller and resource optimizer for the AI industry. Blockchain can limit the overdevelopment and unauthorized operation of AI, solve the problem of data and asset rights confirmation to protect user rights, and integrate and optimize the data and computing power allocation required by AI. However, it is limited to promoting transparency, decentralization and data rights confirmation of AI.
Reference
[1] “Beyond Web3, the Fantastic Journey of AIGC, the New Favorite of Capital” 0xmin (2022.10)
[2] “AIGC Dilemma and Web3’s Way to Break the Circle” wheart.eth (2022.11)
[3] “AIGC: The Revolution of Content Productivity” by Yang Renwen (December 2022)
[4] Zeming Lin (2023.03)
[5] “Brief Thoughts: The Impact of AI Explosion on Creators and NFTs” Sleepy (2023.04)
[6]《Bitcoin: A Peer-to-Peer Electronic Cash System》Satoshi Nakamoto （2009.03）
[7]《Mastering Bitcoin》Andreas Antonopoulos (2016.03)
[8]《Challenges and Recent Advances》Blockchain-Based Payment Channel Networks (2021.07)
[9]《How AI Can Help Build Web3》crypto.com (2023.03)
[10]《Ethereum White Paper》Vitalik Buterin （2023.05)
