Author: CloudY, Jam

Editor: Vincero, YL

Reviewer: Yasmine

In late November 2022, OpenAI launched ChatGPT, an intelligent conversational system, garnering global attention and stimulating extensive discourse.

Equities within the AI sector experienced a notable surge across the A-share market, the U.S. stock market, and the realm of cryptocurrencies. As ChatGPT gained widespread adoption, its profound impact on the global landscape became apparent, leading to the emergence of novel application scenarios and iterative products of similar nature.

Even Microsoft's acquisition of OpenAI and the subsequent integration of ChatGPT into the Bing and Office ecosystem generated soaring investor expectations, reflected in a significant increase in stock price. However, the introduction of ChatGPT4, showcasing superhuman artificial intelligence capabilities, tempered initial excitement and prompted individuals to contemplate the transformative effects of AI on their respective industries and the potential risks associated with further AI advancement.

Against this backdrop, this article aims to explore and address these inquiries by undertaking comprehensive research of both the AI and Blockchain industries, seeking to provide insights and solutions.

Current Development Status of the AI Industry

Productivity Tools

AI can be regarded as a transformative productivity tool, akin to the impact of historical technological advancements such as stone tools, steam engines, internal combustion engines, electric motors, computers, and the internet on human society. By minimizing the barriers to human-computer interaction and augmenting the efficiency of repetitive production tasks, AI can induce substantial shifts in productivity and production relationships. Consequently, AI's influence extends to improving the overall quality of human existence and mitigating impediments to human progress.

AI technology has significantly influenced diverse sectors, including intelligent manufacturing, healthcare, finance, transportation, education, etc. By enabling machines to acquire knowledge and autonomously execute non-creative tasks, AI contributes to enhanced productivity and cost reduction in specific industries. Notably, in pharmaceutical research, AI finds application in protein structure prediction. The ESMFold model was developed by the Meta AI team, which effectively predicted protein structures from a vast dataset comprising over 600 million macro genomes. This remarkable achievement unveils the extensive scope and variety of natural proteins, surpassing previous bounds of the imagination.

In practical terms, AI technology enables the processing complex programs through natural language. It obviates the necessity of comprehending intricate programming or possessing coding proficiency. Instead, users can express their desired outcome to the AI system, which autonomously executes the requisite intermediate steps to attain the intended result. This augmented productivity stems from AI's capability to bridge the divergence between human intentions and task execution, negating the requirement for extensive programming expertise or comprehension of intricate algorithms.

The AIGC technology holds vast potential for applications in various domains, including intelligent customer service, virtual agents, and gaming. By leveraging existing language datasets, ChatGPT enables a seamless and natural conversational experience in virtual agent systems and gaming platforms, enhancing user satisfaction and product competitiveness. Additionally, ChatGPT effectively replaces humans in repetitive content generation tasks, such as generating reports, gathering and summarizing information, translating, and producing conditional illustrations. This AI augmentation liberates human productivity, allowing individuals to focus on providing essential instructions and engaging in creative pursuits, relieving them from mundane task execution.

Technology Trends Guide

The prevailing core applications of AI encompass general Artificial Intelligence, Knowledge Graphs, Data Analysis and Synthesis, Autonomous Driving, and AI-Generated Content (AIGC).

• Knowledge graphs: Graphical representations of diverse entities, relationships, and attributes in knowledge graphs support intelligent search, recommendation, and question-answering applications.

• Synthetic data: Generated through machine learning and other AI techniques, synthetic data is used to train and evaluate AI models, overcoming challenges related to privacy and security when obtaining or sharing real data.

• AIGC: AIGC technology, utilizing deep learning and generative models, is widely discussed and applied in domains such as text generation, audio generation, image generation, video generation, and more.

In 2022, AIGC experienced a significant breakthrough in market financing and media attention. However, it is essential to note that AIGC is still a nascent technology and is in the early stages of exploration and development.

Specifically, the development stages of AIGC can be categorized as follows:

• Research stage: This phase primarily focuses on elucidating the fundamental principles and algorithms of AIGC, investigating methodologies for model training and optimization, and establishing comprehensive databases.

• Application stage: AIGC initiates its deployment in diverse real-world scenarios, exploring effective integration of AIGC technology into specific domains.

• Industrialization stage: AIGC undergoes widespread adoption across multiple industries and fields, giving rise to its distinct industry chain and complementary ecosystem.

In summary, we have recently transitioned from the research stage to the application stage, indicating that the development of AIGC is still in its nascent phase.

Key Components

Data, algorithms, and computing power are the three pivotal factors propelling the advancement of AI.

In the data domain, the increasing significance of data quality and diversity accompanies the ongoing evolution of AI technology. Besides abundant domain-specific data, effective data cleaning, preprocessing, and labeling are imperative for improving algorithm training accuracy. Furthermore, cross-modal and cross-domain data fusion is critical in extracting enhanced value and intelligence.

Concerning algorithms, the current state of AI technology exhibits iterative advancements and continuous refinement. Future trends center around deep learning algorithms encompassing multi-modal and large-scale models alongside innovations in autonomous learning, knowledge transfer, and incremental learning. These developments will elevate the intelligence level and expand the application scope of AI algorithms, facilitating the widespread adoption of AI technology.

Regarding computing power, the acceleration and optimization of AI computations drive ongoing hardware upgrades and enhancements. Specialized chips like GPUs and TPUs have emerged as crucial contributors, significantly amplifying the efficiency and speed of AI computations. Additionally, cloud and edge computing advancements offer more flexible and diverse computational environments for AI processing.

The Current Stage of Blockchain Industry

Distributed Ledger

Blockchain is a decentralized distributed ledger.

Blockchain is a decentralized and distributed ledger with the crucial immutability property derived from its underlying consensus mechanism. On-chain data is recorded in blocks and validated by miners/validators, forming a continuous chain. Once data is recorded in a block, whether generated by smart contracts or accounts, it becomes unalterable.

The difficulty and cost of disrupting consensus increase with the number of nodes, geographic distribution, computational power, or the value of staked tokens. As a result, altering the recorded content becomes a formidable task for centralized entities.

Moreover, in an unalterable setting, smart contracts, constructed through code, empower users to engage with them without relying on any third party for trust. These intelligent contracts execute predetermined code paths to facilitate relevant operations, ultimately enabling the realization of trustless transactions on the blockchain.

Furthermore, assets within the smart contract can only be accessed by the associated account, preventing other accounts from transferring assets from the original account through the smart contract. Each operation of the original account requires a signature to confirm identity, and even the initial transfer interaction requires prior Approve for the smart contract to access the account's assets. This design positions the user's wallet account as the ideal vehicle for their identity (DID) and assets.

Within the framework of consensus mechanisms and smart contracts, all on-chain assets and actions can be recorded and attributed accurately, facilitating the automatic aggregation of related benefits into the rightful owner's account. This effectively resolves the problems of counterfeit assets and impersonation, as it prevents unauthorized individuals from copying and pasting to steal assets or usurp the interests of the rightful owner.

Specifically, digital assets can be uniquely defined using tokenized smart contract addresses. For example, non-fungible tokens (NFTs) can represent digital artworks. Additionally, individuals' actions can be authenticated using non-transferable tokens (SBTs), providing proof of their work or presence in a specific time and space (Proof of Work/Proof of Attendance).

Technology Trends Guide

The layered structure of the Blockchain technology architecture is characterized by Layer 0-2, with consortium chains and private chains representing distinct types of Blockchain application scenarios.

• Layer 0 refers to the physical infrastructure and network architecture of the Blockchain, encompassing hardware devices, network protocols, and transmission media. It serves as a foundational component enabling cross-chain communication and addressing asset-related issues. Notably, leading technologies such as Cosmos, Polkadot, and LayerZero are prominent representatives within this domain.

• Layer 1, also known as the base layer or public chain, plays a fundamental role in the Blockchain ecosystem. Prominent examples of Layer 1 include widely recognized platforms like Bitcoin and Ethereum. The protocols' design and technological implementation at Layer 1 have a significant influence on the core performance and functionalities of the Blockchain system. Furthermore, Layer 1 can be further categorized into distinct types, such as EVM (Ethereum Virtual Machine) and non-EVM-based systems, based on their specific characteristics.

• Layer 2 refers to the protocols and solutions built on top of Layer 1, aiming to enhance the performance and expand the application scenarios of the Blockchain. There are currently six types of Layer 2 protocols, with ZK Rollup and Optimistic Rollup being the mainstream ones. These protocols enable the Blockchain to process a greater number of transactions, improve TPS, and reduce Gas fees.

• A Consortium Chain is a collaborative blockchain network governed by multiple organizations or institutions with shared interests, such as banks, insurance companies, and supply chain companies. It differs from public chains as it has a restricted number of participants and nodes, leading to enhanced transaction speed and security.

• A Private Chain is a permissioned blockchain network belonging to a single organization or institution, usually allowing only internal personnel to participate.



Key Components

Distributed nodes, cryptography, consensus algorithms, smart contracts, and cryptocurrencies constitute the foundational elements propelling the advancement of Blockchain technology.

Distributed nodes constitute the foundational essence of Blockchain technology, facilitating decentralized storage and transmission of data. Cryptography serves as an essential theoretical instrument, ensuring the security and privacy of the Blockchain. Furthermore, consensus algorithms play a pivotal role in establishing distributed consensus within the Blockchain network. Smart contracts, being self-executing computer programs, enable the execution of diverse logical instructions on the Blockchain. Lastly, cryptocurrencies, empowered by encryption techniques, ensure the security and anonymity of transactions.

Through the utilization of distributed nodes, all participants are able to maintain a comprehensive replica of the data, thereby ensuring both transparency and security. The essential technologies within Blockchain, including hash functions, digital signatures, and asymmetric encryption, are cryptographic applications. These technologies play a crucial role in safeguarding data integrity and verifying identities, all while upholding user privacy.

Through the implementation of consensus algorithms including Proof of Work (PoW) and Proof of Stake (PoS), all nodes can achieve unanimous agreement, ensuring data consistency and immutability. Smart contracts facilitate trustless transactions, eliminating the need for intermediaries and thereby enhancing transaction efficiency and security to a certain degree. The emergence of cryptocurrencies like Bitcoin and Ethereum has propelled the widespread adoption and advancement of blockchain technology.

The Intersection of Blockchain and AI

Amidst the AI revolution, it is imperative to reflect on the extent to which AI has transformed Blockchain, as well as the impact of Blockchain's decentralization and trust capabilities on AI.

Firstly, AI, functioning as a productivity tool, possesses the potential to diminish technical barriers, consequently reducing the hurdles within the blockchain industry and enhancing its overall efficiency.

Secondly, AI-powered games and metaverses will liberate themselves from predetermined settings, ushering in fresh narratives and gaming experiences within the blockchain realm.

Blockchain's smart contracts can establish the domains and boundaries within which AI operates or impose limitations on AI's permissions, thereby preventing its unwarranted proliferation.

Furthermore, the decentralization of blockchain can facilitate the sharing and allocation of resources, including the fundamental data and computational power indispensable for training AI models.

Moreover, the authentication capabilities of blockchain can furnish evidence regarding data integrity, identity validation, and ownership rights, thus mitigating conflicts of interest that may emerge from AI applications.

The Significance of AI for Blockchain

Firstly, AI, as a tool, has the potential to lower the barriers to content creation, allowing individuals without technical expertise to express their creativity and produce high-quality content. This encompasses various domains such as NFT creation, game asset development, metaverse modeling, and code generation.

However, the current utilization of AIGC in the NFT field is predominantly limited to generating simple images, lacking fundamental distinctions from traditional Generative Art. To fully leverage the potential of AIGC in the NFT space, further exploration is required to expand the characteristics of NFTs, akin to how Mirror World employs AI to imbue NFTs with a distinct essence.

Secondly, there is a significant reduction in technical barriers related to code development. Code writing encompasses smart contract deployment and hacking or white-hat activities, representing opposite ends of the spectrum. AI can facilitate the deployment of smart contracts through natural language programming, while adversaries can employ AI to analyze contract code and launch attacks. By leveraging AI, it becomes possible to iterate on deployed contract code, fostering internal competition and establishing a more robust and reliable codebase industry-wide. This foundation allows stakeholders to prioritize optimizing blockchain architecture, designing comprehensive projects, and enhancing gameplay, thus fostering innovation at the business level.

Similarly, the simplification of technical barriers by AI enables the widespread application of previously complex operations. Examples include flash loans, optimal mining strategies, and automated yield acquisition, the Judgment of Head Miner Exit Time, all of which can be accomplished by AI. AI possesses the ability to autonomously program, select paths, and directly execute these operations. This parallels the usage of skill cards in the game Yu-Gi-Oh!, where skill cards activate and take effect automatically. This accessibility empowers ordinary users to engage in operations that were previously limited to those with high technical expertise. For instance, capturing MEV typically requires programming an MEV bot. However, when such tasks become achievable by ordinary individuals, profit margins diminish as participation becomes widespread. Consequently, a gas race ensues, where elevated gas fees erode the value of MEV due to the principles of game theory. Ultimately, this leads to reduced profitability and diminished impact of MEV. This phenomenon exemplifies a form of technological development that stimulates industry optimization.

AI will facilitate the widespread adoption of blockchain technology. Currently, there are fewer than 320,000 active Ethereum users, representing a small fraction of total internet users, according to data from Footprint Analytics. The primary challenge lies in the lack of user demand and the complexities of on-chain interactions. Previously, integrating data into the blockchain or using blockchain-based tickets and credentials required establishing a blockchain system or incurring high gas fees, resulting in significant costs. However, leveraging AI technology now allows for low-cost blockchain construction and optimization of on-chain data usage, leading to reduced gas fees. As a result, blockchain technology can be applied and smart contracts deployed in various domains requiring authentication and transparency. Ultimately, an AI-driven simplified interaction system will attract a significant number of users to the blockchain industry.

The impact of AI in the blockchain context is primarily limited to the application layer. Users can leverage AI to bypass the complexities of writing smart contracts and directly deploy applications tailored to their needs. Consequently, project development emphasis will shift from issuance to innovation and operations. The application layer is expected to undergo substantial transformative changes in the future. However, AI's influence does not extend to the underlying layers, including the execution, consensus, and data layers, which necessitate fundamental advancements. Mere automation of repetitive tasks is insufficient to drive qualitative transformations in these areas. For example, the implementation of EIP1559 in the Ethereum London upgrade has bolstered Ethereum's progress, while the completion of the Shanghai upgrade is crucial for increasing ETH staking volume, reinforcing Ethereum's security, and revitalizing the LSD sector's growth.

The Role of Blockchain in AI

The inherent disparity between the decentralization of blockchain and the centralized AI technology paradoxically provides an opportunity to address the challenges encountered in AI.

The predominant centralization of modern AI and big data technologies under the control of a limited number of powerful entities with significant technological capabilities and resources confers influence over market trends and user behavior. Consequently, individuals are compelled to trust AI's faithful execution of instructions, leading to inherent risks such as privacy breaches, algorithmic biases, and data misuse.

The distributed and decentralized nature of blockchain provides a practical solution to these challenges. Through smart contracts, data accessibility and operational boundaries can be limited, thus alleviating the risk of malicious behavior. Deploying monitoring nodes enables penalizing misconduct by confiscating AI's computational resources. This framework ensures AI's directed focus on human development, preventing excessive utilization and unauthorized actions.

Blockchain empowers anonymous users to decide whether to contribute underlying necessary data for AI model training. Zero-knowledge (zk) technology enables the disclosure of user data while preserving personal privacy. The entire process of data collection, storage, and sharing operates on decentralized nodes, ensuring data security, availability, and source verification. Consequently, a proportionate share of the profit generated by AI model can be distributed as dividends to data owners. A suitable incentive mechanism can leverage the decentralized nature of blockchain with high data security.

Likewise, users prompting AI models can also receive partial profits based on their ownership of the prompts when utilized. This arrangement safeguards the interests of both AI data owners and prompt providers.

Computational mining is a crucial consideration due to the substantial data and adequate computational power requirements. However, the current global supply of computational resources falls short of demand. To address this, decentralized cloud computing mining pools can aggregate resources and provide subsidies to contributors. Subsequently, the auctioning of computational power for AI model training ensures efficient utilization of limited resources with computational security and reliability. Furthermore, the integration of data, algorithms, and computational power enables the development of an AI-as-a-Service protocol. Leveraging decentralization and reusability, this protocol offers AI model construction services to users in need, covering data acquisition, processing, algorithm selection, and computational resource allocation. This ecosystem-based approach mitigates centralization risks while preserving supply chain benefits.

In the realm of AI application, blockchain effectively addresses issues like piracy, plagiarism, and virtual identities that arise from AI's remarkable learning capabilities. By recording artworks as on-chain NFTs, unique smart contract addresses validate their authenticity. The value of artworks, besides their inherent artistic qualities, also depends on the identity of their creators, just as imitations of Van Gogh's Sunflower hold little value. While blockchain can prove which sunflower painting is truly crafted by Van Gogh's hand. Also, blockchain can be employed to build distributed knowledge graphs, ensuring data integrity, permanence, and availability.

To address the construction of virtual identities using personal data by AI, Owner Attested Tokens (OATs) or Self-sovereign Biometric Tokens (SBTs) can be utilized. Each blockchain action is logged, and the corresponding OAT or SBT created is distinct, enabling identity verification based on these tokens. Blockchain's tamper-proof nature guarantees the impossibility of fabricating non-existent events.

In summary, AI serves as a productivity tool, propelling the adoption of blockchain and introducing novel narratives for metaverse. However, AI is limited to replacing repetitive tasks and reducing technical barriers, unable to drive innovation in critical technologies. Consequently, AI's impact on the blockchain industry remains confined to the application layer.

On the other hand, blockchain functions as a risk controller and resource optimizer in the AI industry. It curbs excessive AI development and unauthorized operations, safeguards data and asset ownership rights, and optimizes the integration of data and computational resources required by AI. Nonetheless, its purview is primarily concerned with promoting transparency, decentralization, and data ownership within AI.

Reference

• [1]"Bitcoin: A Peer-to-Peer Electronic Cash System" by Satoshi Nakamoto (2009.03)

• [2]"Mastering Bitcoin" by Andreas Antonopoulos (2016.03)

• [3] "Challenges and Recent Advances in Blockchain-Based Payment Channel Networks" (2021.07)

• [4] "Beyond Web3: The Fantastic Drift of AIGC, the New Darling of Capital" by 0xmin (2022.10)

• [5] "The Dilemma of AIGC and the Way to Break Barriers in Web3" by wheart.eth (2022.11)

• [6] "AIGC: The Revolution of Content Productivity" by Yang Renwen (2022.12)

• [7] "Emergence and Evolutionary Information of Large Language Models: Accelerating Protein Structure Prediction" by Zeming Lin (2023.03)

• [8] "How AI Can Help Build Web3" from crypto.com (2023.03)

• [9] "Reflections: The Impact of AI Breakthroughs on Creators and NFTs" by Sleepy (2023.04)

• [10] "Ethereum White Paper" by Vitalik Buterin (2023.05)

#OpenAI #chatGPT-4 #crypto2023 #Binance