Stuart Russell: “Humans love for other human bein…

On the morning of 27 March 2024, Prof. Stuart Russell from University of California at Berkeley shared his insights on AIGC and AI development during the session “AIGC Reshaping the World” at the Boao Forum for Asia (BFA). Prof. Stuart Russell is an expert in artificial intelligence (AI), he is the author of “Artificial Intelligence: A Modern Approach” which is used in over 1500 universities across 135 countries.

Prof. Russell expressed gratitude for the opportunity to participate in this forum. He firstly reminded us about the goal or achievement of developing AI. Over the past 75 years, lots of research have been constructed in various fields. He considered that even if AI has made significant progress, it has not yet succeeded due to technical gaps that cannot be bridged simply by scaling up models or increasing training data.

He emphasised that humans need to keep the control even facing the more powerful technology. As the popular technology Large Language Model (LLM), it likes a black box which making it difficult for humans to comprehend its internal workings and limitation. Prof. Russell thought that is the incorrect direction to go further in researching and emphasised AI models should be transparent and comprehensible. In addition to highlighting the risks associated with AI, he also discussed the potential benefits in supporting education and scientific research.

In terms of future directions, Prof. Russel thought that the next generation of ChatGPT may be released very soon, and its training data may become the ten times of the previous version. However, this may lead to a lack of data. Using synthetic data is not a viable solution as it does not accurately reflect real-world data. He believed that the next milestone should focus on whether AI system can learn and produce products like a human with understanding the inner working of the black box.

Recall to 2023, an open letter was issued requesting a six-month pause in the development of AI to allow time for consideration of how to regulate this technology. This letter served as a reminder to release regulations. A real problem is that AI can be used to solve a lot of problems, it makes the process becomes cheaper and more effective, it can also lead to a loss of motivation in working and thinking. Prof. Russell reminded humans to consider whether this is a right direction and whether it will lead to the better development of human civilisation.

The session discussed the future capabilities of AI, its value to humans and the great challenges in current society. Finally, Prof. Russell summarised the discussion with a sentence “Humans love for other human beings”.

继续滑动看下一个轻触阅读原文 博鳌亚洲论坛向上滑动看下一个

原标题：《Stuart Russell: “Humans love for other human beings”》

阅读原文