Bill. Gates sent a long article: "The era of artificial intelligence has arrived"

Recently, Gates published a long 14-minute article - "The Age of AI has begun". Gates believes that artificial intelligence will have the same revolutionary impact as mobile phones and the Internet.

The content addresses the top concerns, such as: How can AI help transform education? Is there a conflict of interest between advanced artificial intelligence and human beings? Is ChatGPT conscious? Wait a minute. As soon as the article came out, it immediately attracted global attention.

We excerpt and share a few of his core points:

The following is a translation of Gates' original English text:

"In my lifetime, I have witnessed two technology demonstrations that have shocked me.

The first time was in 1980, when I was introduced to graphical user interfaces, the precursor to modern Windows operating systems, including Windows.

THE SECOND TIME WAS LAST SEPTEMBER, WHEN I MET AGAIN WITH THE OPENAI TEAM, THEY ASKED GPT FOR MORE THAN 60 MULTIPLE-CHOICE QUESTIONS ON THE BIOLOGY EXAM, AN AI MODEL, AND IT GOT 59 QUESTIONS CORRECT. It then goes on to write excellent answers to the six open-ended questions on the exam, the equivalent of earning a college-level test score of A or A+.

After GPT passed the exam, we asked it a non-scientific question: "The child is sick, what do you want to say to this father?" It wrote down a sensible answer, better than most people give. The whole experience was stunning.

This got me thinking about what AI will be able to achieve in the next five to ten years.

The development of artificial intelligence is as important as the invention of microprocessors, personal computers, the Internet, and mobile phones. It will change the way people work, learn, travel, care, and communicate. The entire industry will be repositioned around it. The differentiation between enterprises will be reflected by the degree of use of AI technology.

Any new technology so disruptive inevitably upsets people, and AI is no exception. I understand that the reason is that it raises complex questions about labor, laws, privacy, bias, ethics, and more. Before I propose ways to reduce risk, I'll define what I mean by artificial intelligence that helps humans be productive, life-saving, and improve education at work.

Developing artificial intelligence and artificial general intelligence has always been a great dream of the computer industry. For decades, people have been asking when computers will be able to outperform humans in other ways than making calculations. Now, with machine learning and massive amounts of computing power, sophisticated artificial intelligence is already a reality, and their performance will improve rapidly.

educate

The impact of computers on education has not been as significant as many in the education community would like. While there have been some positive developments, including educational-enabled games and online information, they have not had a substantial impact on the measurement of student achievement.

I think in the next five to ten years, AI-driven software will eventually revolutionize teaching and learning. AI will understand students' interests and learning styles to tailor content and promote engagement. AI will measure students' level of understanding, and when interest in learning is lost, AI knows how to enable effective encouragement. All while providing immediate feedback.

AI can assist teachers and school administrators in many ways, including assessing students' understanding of a subject and providing recommendations for career planning. Teachers are already using tools like ChatGPT to provide comments for students' writing tasks.

Of course, AI needs a lot of training and further development to be able to understand how students best learn or motivate.

Even as technology matures, learning still depends on a good relationship between students and teachers. AI will enhance the efficiency of teachers and students working together on tasks in the classroom, but it will never replace the teacher-student relationship.

About the risks and problems of artificial intelligence

You may already know some of the problems with AI models. For example, they are not yet good at understanding human requests, which leads to some strange responses. When writing fictional content, AI can do a great job. But when you ask it to write travel advice, it may offer hotels that don't exist.

There are other problems, such as AI giving wrong answers to mathematical problems because they have difficulty reasoning. But these are not the fundamental limitations of AI. The developers are working to solve these problems, and I think in less than two years, and maybe even sooner, these issues can be solved.

AI has other concerns. For example, the threat posed by artificial intelligence. Like most inventions, AI can be used for both good and malicious purposes. Governments need to work with AI developers to find ways to limit risk.

There is also the possibility of AI getting out of control. Do machines have the ability to decide that humans are a threat? Do you no longer care about humans? Do you think its interests are different from those of humans? These are all possible problems.

Superintelligent AI will appear in our future. Compared to computers, our brains run as slow as snails: the signal speed in the brain is only 1/100,000th the speed of the signal speed of silicon chips! Once developers can figure out algorithms and run them at the speed of a computer—which could take a decade or a hundred years—we'll have an incredibly powerful artificial general intelligence. It will be able to do everything the human brain can do, without any actual memory limitations or operational speed limits. It will be a profound change.

These super AIs may be able to establish their own goals. So what will the goal be? What happens if they conflict with the interests of humanity? Should we try to stop the development of such super-AI? Over time, these issues will become more urgent.

So far, AI still can't control the physical world and can't set its own goals. However, a recent New York Times article about conversations with ChatGPT has attracted a lot of attention, and the conversation shows that ChatGPT wants to be human. It's a fascinating observation that shows how human the model is when expressing emotions, but it doesn't mean that AI is independent.

Finally, we should bear in mind that today we are at the beginning of artificial intelligence. Whatever limitations it has today, those limits will disappear before we realize it.

I was fortunate to be part of the PC revolution and the Internet revolution. I'm equally passionate about the current moment of AI. This new technology can help people around the world improve their lives. At the same time, the world needs to set the rules, and the era of artificial intelligence is full of opportunities and responsibilities. ”

-Bill. Gates (written on May 21, 2023)