How Artificial Intelligence "Towards Good"

Tech 2023-06-19 08:40:15 Source: Network

For a period of time, the artificial intelligence big model represented by ChatGPT has stirred up the wave of global artificial intelligence technology development. From writing code to telling stories, from writing articles to automatically creating data tables

For a period of time, the artificial intelligence big model represented by ChatGPT has stirred up the wave of global artificial intelligence technology development. From writing code to telling stories, from writing articles to automatically creating data tables... Artificial intelligence is bringing many changes to human work, learning, and life.

How far are we from the "omnipotent" universal artificial intelligence? What security risks and challenges does the development of artificial intelligence bring? At the recently held 2023 Beijing Zhiyuan Conference, artificial intelligence experts and scholars from around the world discussed related topics.

Universal artificial intelligence has a long way to go

"Imagine that in the next 10 years, General Artificial Intelligence (AGI) will surpass human expertise in almost every field, and may eventually surpass the overall productivity of all large companies, which will improve people's living standards." Sam Altman, CEO of OpenAI, presented a vision of the future of AI.

The so-called AGI refers to an artificial intelligence system that can perform intelligent tasks in various fields like humans. This is different from the current AI application which only focuses on specific tasks or fields (such as image recognition, speech recognition, Natural language processing, etc.), which puts forward higher requirements for AI technology.

"General artificial intelligence can learn and perform tasks better and faster than human beings, including tasks that human beings cannot handle. Due to the huge advantages of machines in speed, memory, communication and bandwidth, general artificial intelligence will far exceed human capabilities in almost all fields in the future." said Stuart Russell, a professor of computer science at the University of California, Berkeley.

Although artificial intelligence has a "timeline" that surpasses humans, in the eyes of many experts, there is still a considerable distance between current artificial intelligence and AGI.

Russell believes that the current popular big language model does not "understand the world", but is just a "puzzle" of universal artificial intelligence - "We do not know how to connect it with other parts, and there are even some missing puzzles that have not been found

Huang Tiejun, president of Beijing Zhiyuan Artificial Intelligence Research Institute, pointed out that there are three Technology roadmap to achieve general artificial intelligence: first, large models, which enable artificial intelligence to have the ability to emerge through massive high-quality data; The second is embodied intelligence, through Reinforcement learning method, training to produce body model; The third is brain like intelligence, which enables machines to achieve or resemble human brain abilities.

For the development of artificial intelligence, Yann LeCun, a Turing Award winner and professor of New York University, put forward the concept of "world model" - through this model, artificial intelligence systems can understand the way the world works, and act in an optimized and least cost way.

Strengthening international cooperation in the field of security governance

According to PwC, artificial intelligence will create an economic value of $15.7 trillion by 2030. Artificial intelligence provides important opportunities for economic development, but it also raises concerns and controversies about security.

Geoffrey Sinton, a Turing Award winner and professor at the University of Toronto, believes that the current AI can master the way to "cheat" humans through learning. Once artificial intelligence has the ability to deceive, it has the ability to control humans. Such super intelligence may occur faster than expected

Before the arrival of the era of universal artificial intelligence, the security risks of artificial intelligence mainly came from "people". "We should not assume that machines are fair, because machines may try to change human behavior. More accurately, it is the owners of machines who want to change the behavior of others." Andrew Yao, winner of the Turing Award and academician of the CAS Member, said that the development of artificial intelligence is currently in an important window period, and countries should work together to build a governance structure for artificial intelligence.

With the increasing power of artificial intelligence, the issue of "alignment" in artificial intelligence has emerged. The so-called "alignment" refers to the alignment and consistency between the goals of artificial intelligence systems and human values and interests.

How to align artificial intelligence with humans? Altman believes that people should responsibly apply AI to the world and pay attention to and manage security risks. He suggests establishing equal and unified international norms and standards in the development process of artificial intelligence technology, and establishing a trust system for the secure development of artificial intelligence systems in a verifiable manner through international cooperation.

Huang Tiejun believes that although artificial intelligence may generate unexpected new capabilities, this does not mean that humans cannot manage artificial intelligence. How to manage a highly creative system like artificial intelligence can provide great reference significance for disciplines such as sociology and history

In February of this year, China proposed in the Global Security Initiative Concept Paper to strengthen international security governance in emerging technology fields such as artificial intelligence, prevent and control potential security risks. At this Smart Source Conference, experts and scholars actively evaluated China's contribution to promoting international governance in artificial intelligence.

Altman said that China has a large number of excellent talents and product systems in the field of artificial intelligence, which should play a key role in the safety of artificial intelligence.

Max Tegmark, professor of the Center for Artificial Intelligence and Basic Interaction Research at MIT, said that China's growing ability to shape the global AI agenda can play a leading role in the field of AI security governance.

Promote the co construction and sharing of large models

Currently, the global technology competition in the field of artificial intelligence is becoming increasingly intense. The "Research Report on China's Artificial Intelligence Large Model Map" released at the 2023 Zhongguancun Forum shows that 79 artificial intelligence large models with parameters above 1 billion have been released nationwide.

From a global perspective, the number of large models released by China and the United States exceeds 80% of the global total. Since 2020, China has entered a period of rapid development of large models, and has established a systematic research and development capability covering theoretical methods and software and hardware technologies in the field of large models. It has formed a group of large model technologies that closely follow the world's forefront, and has emerged with multiple pre trained large models with industry influence.

Huang Tiejun believes that the large model of artificial intelligence has three characteristics: firstly, it has a large scale; Secondly, there is' emergence ', which means that unexpected new abilities can be generated; The third is universality, not limited to solving specialized problems or specialized fields. He stated that the big model is not a technology monopolized by any institution or company, and should be jointly built and shared to introduce a basic algorithm system required for an intelligent society.

(Intern Zhang Weina also contributed to this article)

Disclaimer: The content of this article is sourced from the internet. The copyright of the text, images, and other materials belongs to the original author. The platform reprints the materials for the purpose of conveying more information. The content of the article is for reference and learning only, and should not be used for commercial purposes. If it infringes on your legitimate rights and interests, please contact us promptly and we will handle it as soon as possible! We respect copyright and are committed to protecting it. Thank you for sharing.(Email:[email protected])

Mobile advertising space rental

Tag: How Artificial Intelligence Towards Good