Microsoft poured $ 1 billion in OpenAI investment, research on super-AI but “ethical”

Tram Ho

Microsoft is booming, and with its profits, they are looking for ways to invest to continue earning momentum.

One of Microsoft’s most notable investments recently is the $ 1 billion drop in startup startup OpenAI artificial intelligence. The goal of Microsoft and OpenAI is to create artificial intelligence that is not only strong in research capabilities, but also reliable and “ethical”.

Đang tải Tinhte_AI3.jpg…
Founded in 2015, OpenAI is one of the few companies that focuses on AGI development (Artificial General Intelligence – I temporarily translate it as a Artificial Intelligence), another party who is also studying AGI is Google DeepMind. According to OpenAI itself, “AGI will be a system capable of doing more in-depth research than any other person in the world. It is similar to the intellectual tool of Marie Curie, Alan Turing and Johann Sebastian Bach.

An AGI system that studies how to solve a problem will be able to connect facts that people cannot do. We want AGI to work with people to solve current multidisciplinary issues, including global challenges such as climate change, high quality medical services for everyone, and teachers. personalized education. We think the impact of AGI will create a free economy for everyone, allowing them to pursue what they are passionate about, creating new opportunities that are unimaginable today. ”

Đang tải Tinhte_AI1.jpg…
The words describe the ambitious OpenAI vision. Currently they themselves are having breakthroughs in artificial intelligence that not many parties can do. The first is the “reinforcement learning” ability, self-learning enhancement, OpenAI 5 has defeated the people in DOTA 2. This project not only describes the potential of AI in video games, but also Describe the advantage of self-reinforcing learning for many other industries such as robotics, retailing or manufacturing assembly. The second is the advances in NLP array, natural language processing. As a result, they can create systems that meet the needs of human language interactions such as translation, summarizing, or even creating texts like people write.

Đang tải Tinhte_AI4.jpg…

It is also impossible to ignore the fact that at the present time AI is still not as strong as many people think, the application in many industries is not high. The same is true for AGI, it must be applicable for a while. Even the AI ​​doesn’t know how to learn to create a worldview like animals learn for hundreds of thousands of years. It is true that AI can defeat people in some areas thanks to analyzing a lot of data. But people do not understand and interact with such a world. Instead, we make predictions and use arguments to analyze the future from existing data. We can extrapolate events with normal thinking and work on incomplete topics. That’s what AI hasn’t done yet. In short, easy to understand, we can not create an autonomous AI to serve the camera scanning system, the computer can understand all the languages ​​in the world.

Đang tải Tinhte_AI2.jpg…
That was when OpenAI and Microsoft had to study more years to solve this difficult challenge, and when they succeeded, both OpenAI and Microsoft would have advantages and profits in the near future. Not to mention the “moral” issue of AI. When artificial intelligence knows how to dig for data themselves, people need to be more concerned about the impact of AI on society. Deepfake is a good example.

Share the news now

Source : ITZone via TinhTe