Should we be happy or worried about AI?

Leading scientists from home and abroad shared and found answers to the hottest topics about artificial intelligence (AI) at the VinFuture Foundation's discussion on the afternoon of December 4.

Professor Yann Lecun, New York University and Director of Artificial Intelligence Science at Meta, USA - Photo: M.THANH

"You haven't seen it yet, but AI will get better in many years to come. Current AI still has limitations. To bring AI closer to "Human intelligence in the future could be 10 years from now...", Professor Yann Lecun, New York University and Director of Artificial Intelligence Science at Meta, USA, who is known as the "father" of AI, emphasized in his opening presentation at the talk show "Implementing AI in practice".

The seminar is part of a series of events of the Science and Technology Week and the VinFuture Awards Ceremony 2024. Gathering world-famous names in the field of AI, this is a forum for scientists to discuss the most controversial topics today: Reality and misconceptions about artificial intelligence (AI), how do large AI models impact society? What management policies and tools are needed to manage and minimize risks from AI? Predicting the next steps of AI development and strategic orientation for Vietnam in the new context.

The panel discussion "Implementing AI in practice" attracted the attention of many researchers, academics and businesses - Photo: M.THANH

Seize the opportunity from AI, not fear it?

Professor Yann Lecun heated up the discussion by stating that we should not be afraid of the development and impact of AI. He said that we should expand the scale of AI applications to create practical impacts. When AI has human-level intelligence, it will be like having an assistant to support us in our work, "we should take advantage of the opportunities and we should not be afraid", he emphasized.

"We have countless different opinions about AI, both positive and negative. Some people are worried that it will replace us, but others see that computer systems with AI are better, completing millions of calculations, better than humans. We need to see which opinion prevails," said Professor Yann Lecun.

"In the 60s, people said that AI would replace humans in 10 years, but now it is just developing. There are still many difficulties, but we cannot imagine them all. Therefore, it is very difficult to have a direction for AI development," the professor added.

"Let's abandon the generative, probabilistic or supervised learning AI models and approach them in a smarter way. Let's find a way to supplement human knowledge, let's make the AI ​​platform open, and let's have a system in the future that can speak every language in the world, if AI can be trained," Professor Yann Lecun suggested to guide AI development.

Because according to him, "it is necessary to understand that if there is an entity that is smarter than us, it does not mean in all fields but only in a few fields. AI needs to develop, it needs humans to teach skills".

Leading AI scientists from the United States and Vietnam share their views on the future of AI - Photo: M. THANH

Sharing this view, Professor Leslie Grabiel Valliant (Harvard University, Member of the VinFuture Prize Council) also said: "We understand very well the development process of AI and technology and we can agree on one common point which is the importance and significance of AI".

According to him, "science and technology is a vast field, so how to use AI is still a big problem. We can use it in different ways. But we must agree to see the potential of AI in the future. There are still people who do not completely agree, do not agree on the superiority of AI, but we can see what is being built by AI will help to unify the views more."

How can AI be developed with limited resources?

According to Dr. Bui Hai Hung, General Director of VinAI, the cost of investing in AI research and development is huge, requiring hundreds of millions of USD for one project.

"I asked the question: is this sustainable? What about insufficient resources when the hardware is expensive and uses a lot of energy? What are the issues related to maintenance? What is the most important issue when deploying AI in Vietnam?", he raised the issue and also gave the answer: "Relatively limited resources are something to face. In 2019, when I returned to Vietnam from the US to work at VinAI, I realized that affordability and access are very important, but the way is to turn limited resources into a catalyst for innovation."

Dr. Bui Hai Hung, General Director of VinAI and Dr. Xuedong Huang, Chief Technology Officer of Zoom Corporation (USA) - Photo: M. THANH

Dr. Bui Hai Hung cited the fact that in five years, Vin AI has had more than 170 articles, trained more than 100 AI interns to participate in doctoral research in the US, Google Deepmind... Mr. Hung said that Vietnam in general and VinAI in particular, are facing the problem that needs to be solved: the ability to access and pay appropriate costs to bring AI to everyone, how to cut costs to make AI models more effective to solve both challenges in AI development: latency and performance.

Responding to the question of what Vietnam needs to improve to build a good research environment for AI, Professor Yann Lecun said: "In Vietnam, universities focus on talent, attract AI research activities there, and motivate young people. Vietnam has the advantage of a young population, investment in higher education, stem industries... will create opportunities for young people to have opportunities, ambitions and connections, and others who have studied abroad will gather here."

The discussion was lively with many questions from the attendees for the scientists _ Photo: M.THANH

"I have been working with colleagues and students for many years in the field of AI. This is a very special period...", Dr. Bui Hai Hung shared. "We can imagine that AI will develop further at some point and I see young people are very excited to see that AI can do more in the future. And we will see very rapid progress in the future".

Going into more detail, Professor Do Ngoc Minh (University of Illinois at Urbana - Champaign, USA and VinUni) suggested: "Vietnam has a large population and diverse medical conditions. Using AI to collect medical health conditions is an area that can be exploited and used."

Dr. Xuedong Huang, Chief Technology Officer of Zoom Corporation (USA), member of the VinFuture Award Council, said that "AI can help hundreds of young developers, other small startups, create a fair playing field, collaborate, discover new ideas... which makes the difference".

Source: https://tuoitre.vn/nen-vui-mung-hay-lo-lang-truoc-ai-20241204180428449.htm