Click the blue text
Follow us
The three fundamental elements of artificial intelligence: data, computing power, and algorithms are interdependent and mutually supportive, jointly promoting the rapid development of artificial intelligence. This article will outline the current development status of cutting-edge artificial intelligence from the perspective of algorithms and models.
｜Current Status of Algorithm and Model Development
Algorithms, as one of the core elements of artificial intelligence industry development, refer to the finite, abstract, effective, and composite control structures that are enforced under certain rules to achieve specific goals. They manifest in various forms such as neural networks, convolutional neural networks, machine learning, and deep learning. For computers, algorithms are the principles and rules for processing information. Ideally, algorithms can formalize human thought processes into a format that can be continuously executed by computers to achieve set goals. On November 30, 2022, AI large models represented by ChatGPT emerged as a new form of algorithmic “work” that gained explosive popularity, ushering algorithms into the era of AI large models. Within just one week of its release, ChatGPT had over 1 million users, and by the end of January 2023, just two months after its launch, its monthly active users had surpassed 100 million. As the fastest-growing consumer application in history, ChatGPT has become a phenomenal product that is popular worldwide.
The Transformer architecture has ushered in the era of rapid development of algorithm models, with multi-modal general AI large models becoming a development trend.In 2017, Google proposed the revolutionary neural network structure based on self-attention mechanisms—the Transformer architecture, laying the foundation for the pre-training algorithm architecture of large models. In 2018, the release of GPT-1 by OpenAI and BERT by Google marked the mainstream emergence of pre-trained large models in the field of natural language processing. The new neural network architecture represented by Transformer established the algorithmic framework for large models, opening a new era of large model development, evolving through single-language pre-trained models, multi-language pre-trained models, and multi-modal pre-trained models, with multi-modal general AI large models becoming the mainstream development trend.China’s AI large model quantity and the number of research and development entities are considerable, providing new momentum for market growth.In May 2023, the New Generation Artificial Intelligence Development Research Center of the Ministry of Science and Technology released the “Research Report on China’s Artificial Intelligence Large Model Map,” which shows that among the globally released cognitive large models, the United States and China account for over 80%. China ranks second in the number of large models developed, with more than half of these models being open source. Liu Liehong, director of the National Data Bureau, revealed at the China Development Forum 2024 that as of March 25, 2024, China has over 100 large models with more than 1 billion parameters, with large technology companies, research institutes, and startup tech teams becoming the main force in large model research and development. On April 2, 2024, the National Internet Information Office released the “Information on Registered Generative Artificial Intelligence Services,” announcing that as of March 2024, 117 large models in China have successfully registered. AI large models are transitioning from specialized AI workshops to large-scale industrial production of general AI, continuously deepening their implementation in vertical application fields, thus providing new momentum for market growth..
(Excerpt from the book “Cutting-Edge Artificial Intelligence: Development and Governance,” with omissions)
“Cutting-Edge Artificial Intelligence: Development and Governance“
Chief Editor: Liang ZhengChina Development Publishing House
Book Introduction
The book “Cutting-Edge Artificial Intelligence: Development and Governance” is advised by Xue Lan, a senior professor and doctoral supervisor from Tsinghua University, and the dean of the Schwarzman College at Tsinghua University. Liang Zheng, a professor at the School of Public Administration at Tsinghua University and deputy director of the International Governance Research Institute for Artificial Intelligence, serves as the chief editor. This book presents the latest development dynamics of China’s AI large models based on the interpretation of the technical principles and functional characteristics of cutting-edge artificial intelligence, showcasing the value prospects of “Artificial Intelligence + Multi-Scenario” through vivid cases, identifying the risks and challenges arising in the development, deployment, and application of technology, and comparing the AI governance models of the United States, the European Union, and other economies. It deeply analyzes the current status and challenges of international governance of artificial intelligence and proposes the development direction and future paradigm for constructing a governance system for cutting-edge artificial intelligence, striving to provide strong support and beneficial references for the healthy development and effective governance of artificial intelligence.
Table of Contents (Scroll up to read)
Chapter 1 Understanding Cutting-Edge Artificial Intelligence1. Cutting-Edge Artificial Intelligence: From ChatGPT to Sora(1) The Origin and Development of Artificial Intelligence(2) The Connotation of Cutting-Edge Artificial Intelligence2. Characteristics and Functions of Cutting-Edge Artificial Intelligence(1) Characteristics of Cutting-Edge Artificial Intelligence(2) Functions of Cutting-Edge Artificial Intelligence3. Advantages and Limitations of Cutting-Edge Artificial Intelligence(1) Advantages of Cutting-Edge Artificial Intelligence(2) Limitations of Cutting-Edge Artificial IntelligenceChapter 2 Development and Application of Cutting-Edge Artificial Intelligence1. Current Development Status of Cutting-Edge Artificial Intelligence(1) Current Status of Data Elements(2) Current Status of Computing Power Elements(3) Current Status of Algorithms and Models(4) Cutting-Edge Artificial Intelligence: Development and Governance2. Overview of the Industrial Ecology of Cutting-Edge Artificial Intelligence(1) Current Status of Industrial Development of Cutting-Edge Artificial Intelligence(2) Current Status of Commercial Deployment of Cutting-Edge Artificial Intelligence(3) Trends in Industrial Transformation of Cutting-Edge Artificial Intelligence3. Application Prospects of Cutting-Edge Artificial Intelligence(1) Functional Application Scenarios(2) Horizontal Field Application Scenarios(3) Vertical Field Application Scenarios4. Industrial Practices of “Artificial Intelligence +” Actions(1) Artificial Intelligence + Industry(2) Artificial Intelligence + Government(3) Artificial Intelligence + Education(4) Artificial Intelligence + Finance(5) Artificial Intelligence + HealthcareChapter 3 Potential Risks of Cutting-Edge Artificial Intelligence…1. Standards for Risk Classification2. Risks Based on the Technology Itself(1) Data Privacy Leakage(2) Algorithm Discrimination and Bias(3) Content Errors3. Risks Based on Technology Development(1) Data Ownership Issues(2) Psychological Issues(3) Environmental Issues4. Risks Based on Technology Application(1) Digital Divide and Social Inequality(2) Intellectual Property Disputes(3) Deep Forgery(4) Ideological Infiltration(5) Risks of Losing Control(6) Threats to National SecurityChapter 4 International Governance of Cutting-Edge Artificial Intelligence1. International Comparison of Governance Models for Cutting-Edge Artificial Intelligence(1) Governance Model of Cutting-Edge Artificial Intelligence in the United States(2) Governance Model of Cutting-Edge Artificial Intelligence in Europe(3) Governance Models of Other Countries(4) Insights for China’s AI Governance2. Building an International Governance System for Cutting-Edge Artificial Intelligence(1) Forming a Basis: International Consensus on Potential Risks of Cutting-Edge Artificial Intelligence(2) Facing Challenges: International Discrepancies in the Development and Governance of Cutting-Edge Artificial Intelligence(3) Entry Points for Establishing an International Governance System: Building an Open and Inclusive International Platform(4) China’s Participation in the Construction of the International Governance System for Cutting-Edge Artificial Intelligence: Development and GovernanceChapter 5 Building a Governance System for Cutting-Edge Artificial Intelligence1. Concepts and Logic—How to Understand AI Governance(1) Definition of AI Governance(2) Multi-level and Multi-perspective of AI Governance2. Collaboration and Agility—How to Build an AI Governance System(1) Governance Dilemmas Facing National AI Governance(2) Walking on Two Legs: Balancing Development and Governance(3) Collaborative Actions: Promoting Participation of Multiple Entities in Governance(4) Model Construction: Promoting Improved Agile Governance3. History and Transcendence—Responding to New Changes in AI Governance(1) Paradigm Shift in China’s Adaptive Governance of AI(2) Paradigm Transcendence in China’s AI GovernanceAppendix: Review of Major AI Events in 20231. China Large Model Section2. China AI Policy Section3. International Governance Section4. Tech Giants Section
About Us
Tsinghua University Institute for AI International Governance (Institute for AI International Governance, Tsinghua University, THU I-AIIG) is a university-level research institution established by Tsinghua University in April 2020. Relying on Tsinghua University’s existing accumulation and interdisciplinary advantages in artificial intelligence and international governance, the institute conducts research on major theoretical issues and policy needs of international governance of artificial intelligence, aiming to enhance Tsinghua’s global academic influence and policy leadership in this field, and to provide intellectual support for China’s active participation in international governance of artificial intelligence.
Sina Weibo: @Tsinghua University Institute for AI International Governance
WeChat Video Account: THU-AIIG
Bilibili: Tsinghua University AIIG