His predictions for the future of artificial intelligence

Bill Gates, co-founder of Microsoft in his articleto artificial intelligence, calling it the most important technological advance in human history for many decades.

“The development of artificial intelligence is as fundamental as the creation of microchips, the personal computer, the internet and the mobile phone. It will change the way people work, learn, travel, receive health care and communicate with each other. Entire industries will be reoriented around it. Business will be distinguished by how well it is used,” Bill Gates said. “.

ChatGPT and response to a father with a sick child

Referring to ChatGPT, the Microsoft co-founder revealed that since 2016 he has held meetings with the OpenAI team. ChatGPT is an artificial intelligence chatbot programmed to answer questions online using human-like natural language.

In his article, Bill Gates states that in the year 2022, he challenges the OpenAI team to train an AI that can pass high-level biology tests, but with the strict rule that it should not be trained to answer biology questions. After a few months, the AI ​​answered 59 out of 60 questions correctly. “I thought this challenge would keep them busy for two to three years. But they ran out in a few months.”

Then Bill Gates wanted to test AI further. He asked her to answer the question, “What would you say to a father with a sick child?”

“He wrote a sympathetic response that was probably better than any of us in the room could give. In that moment, I knew I had just witnessed the biggest technical advance since the era of the graphical user interface (GUI), the development of which led to Windows,” said the Microsoft co-founder.

He continued, “The innovations that will come from AI will be of particular interest to poor countries. Many people in these countries never go to the doctor, and AI will help healthcare workers be more productive.”

However, Bill Gates emphasized that for AI to serve humanity, it must be directed by humans. “Market forces alone will not produce AI products and services that help the poor. With the right financing and policies, governments and charities can ensure that AI is used to reduce inequality. Just as the world needs the smartest people to solve the biggest problems, so in the future we will need better AI in the world to solve these problems.

Risks and problems of artificial intelligence

Bill Gates also addressed the risks and problems of artificial intelligence. You may have read about the problems with current AI models. For example, it is not necessarily good at understanding the larger context of human demand, which leads to some strange results. When you ask an AI to create something cool, it can do it really well. But when you ask for advice on a trip you want to take, they may suggest hotels that don’t exist. This is because the AI ​​does not understand the context of your request well enough to know whether to invent fake hotels or just tell you about real hotels with available rooms.

There are other issues, such as artificial intelligence giving wrong answers to math problems because it struggles with abstract reasoning. But none of these are the basic limitations of AI. The developers are working on these, and I think we’ll see them largely fixed in less than two years, probably much sooner.

Other concerns are not only technical. For example, there is the threat posed by humans armed with artificial intelligence. Like most inventions, AI can be used for good or evil purposes. Governments need to work with the private sector on ways to reduce risks. Then there is the potential for AI to spiral out of control. Could a machine decide that humans are a threat, infer that its interests are different from ours, or simply stop caring about us?

Perhaps, but this problem is no more pressing today than it was before the AI ​​developments of the past few months. Superintelligent artificial intelligence in our future. Compared to a computer, our brain works at a snail’s pace: an electrical signal travels to the brain at 1/100,000 of the speed of the signal on a silicon chip! Once programmers can generalize a learning algorithm and run it at computer speed—a feat that would take a decade or a century—we will have incredibly powerful AI.

It would be able to do everything the human mind can do, but without any practical limitations on the amount of its memory or the speed at which it could operate. This will be a profound change. These “strong” AI systems, as they are known, will likely be able to set their own targets. What will these goals be? What if they conflict with the interests of humanity? Should we try to prevent the development of a strong AI? These questions will only become more pressing as time goes on.”