Neural networks have become quite popular in recent months. This is largely thanks to ChatGPT and generative neural networks. But the developers look to the future and foresee new problems in this area. The OpenAI blog published an article about the fairly rapid emergence of Artificial general intelligence (AGI). This can be translated as “general purpose artificial intelligence”, the authors themselves mean by the term artificial intelligence that can perform any task no worse, and sometimes even better, than a human.Created by Midjourney Neural Network

The article states that AGI should “enhance humanity’s opportunities for maximum well-being,” but it can also bring potential problems. To prepare for these ahead of time, OpenAI proposed a set of principles to guide the team in their work on AGI.

Their list looks like this:

• Deploying and studying AI systems that are not very powerful in the real world allows you to gain experience and feedback with them.
• Simplifying access to AI systems for everyone.
• AI support of human values ​​and preferences.
• Interaction of international structures for the use and regulation of AI.
• Sharing data and experiences on the advancement of AI with the public, as well as involving stakeholders who can contribute to the development of AI.

These are just the main points, others may arise as AGI develops and will depend on the factors listed above. After all, it is still difficult to predict how quickly such an artificial intelligence will develop, how well the coordination of efforts and the interaction between different countries will be. It’s also unclear how exactly AGI will take into account human goals, preferences, and preferences.

OpenAI also noted that they themselves do not have all the answers and solutions to potential future problems, especially after AGI has become a mainstream phenomenon. They expect these and other requirements for the development of artificial intelligence to become mandatory for everyone in the long term.