Shulou(Shulou.com)11/24 Report--

Three Turing Award winners, Hinton, Bengio and Yao Qizhi, jointly wrote "Managing AI risk in an era of Rapid Development".

It has also been co-authored by UC Berkeley, Song Xiaodong, Pieter Abbeel, Tsinghua Zhang Yaqin and other well-known scholars.

This is a consensus paper (consensus paper). In addition to AI researchers, there are also scholars in public governance, such as Xue Lan, Institute of International Governance of artificial Intelligence at Tsinghua University, and Daniel Kahneman, a Nobel laureate in economics.

Hinton believes that companies are planning to increase the computing scale of AI models by a factor of 100 within 18 months, and no one knows how powerful these models will be or how they will use them. (100x of the 18-month plan comes from InflectionAI)

The managing-ai-risks.com,arXiv version of the current paper published on the independent domain name site will be uploaded later.

Looking back at 2019, GPT-2 can't count perfectly.

In just four years, deep learning technology has helped develop software, create realistic virtual environments, provide professional advice, and equip robots with language understanding.

The progress of these systems has exceeded expectations and is eye-popping.

And the next progress may be even more amazing.

Although there are still many limitations in the existing deep learning system, companies are already racing to develop a general AI that can compete with human intelligence. They put in more and more resources, innovative technologies are changing with each passing day, and the self-evolution ability of AI accelerates this trend.

There is no sign that AI's progress will stagnate after reaching the level of human intelligence.

In fact, AI has surpassed human beings in certain areas. They can learn faster, process more information, perform well in large-scale computing, and can be easily copied.

The rapid progress of technology is shocking. Some tech giants have the ability to increase their AI training several times in a short period of time. Given the continued investment in AI research and development and the trend of self-evolution, the possibility that GM AI will surpass humans in the next 10 to 20 years cannot be ignored.

If applied correctly and fairly, advanced AI can solve long-standing problems for mankind, such as disease, poverty and environmental problems. But at the same time, the powerful AI also brings huge risks that we are far from ready to deal with. We need to balance the capabilities and security of AI.

Unfortunately, we have lagged behind in adjusting our strategy. We need to anticipate and deal with risks in advance, and we can't wait until problems really arise. Environmental issues are a case in point, but the speed of AI development means we don't have much time to wait.

The rapid progress of AI has brought great challenges to the society. If left unregulated, they can exacerbate social inequality, threaten social stability and weaken our perception of the real world. They may even be used as tools of crime or exacerbate global inequalities and conflicts.

More worryingly, the development of autonomous AI may amplify these risks. Although the current AI is not fully autonomous, this state is changing. For example, GPT-4, powerful as it is, may become unpredictable and difficult for us to intervene if it controls robots or programs with autonomous goals.

To this end, we call on the world to make joint efforts to regulate the development and application of independent AI through technical, policy and legal means. Otherwise, we may face a future dominated and misused by AI.

This is just the beginning. We urgently need more research and discussion to solve these problems, and we hope to have the support and assistance of the community to ensure that AI brings real benefits to mankind, not threats.

Reference link:

[1] managing-ai-risks.com

[2] https://twitter.com/geoffreyhinton/status/1717967329202491707

This article is from the official account of Wechat: quantum bit (ID:QbitAI), author: Mengchen

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.