ChatGPT is fun, but what if things get serious and AI makes our lives as we know them worse? Experts expect that within 30 to 40 years a superintelligence AI will emerge that will have unprecedented levels of power. A guest post by Christoph Herr.

Yesam Altman, CEO of Open AI, warned at a US Congressional hearing in the spring of this year that artificial intelligence (AI) is a “Danger of Extinction” represents for humanity. He looked very worried and said with a serious expression: “If this goes wrong, then it can be go completely wrong.”

Anyone currently working with tools like ChatGPT is initially happy with the huge relief that comes with it. Knowledge becomes easily and quickly accessible, code can be generated automatically, protocols are created by AI colleagues without human help, and much more. What about that could end humanity?

keep reading

The reason for Altmann’s concern is that today we only see a snapshot of AI’s potential capabilities. Development is far from over and the next steps have already been mapped out. Technology thought leader The following stages of development are expected:

• Level 1: Artificial Narrow Intelligence (ANI): Automation of daily tasks and processes.
• Level 2: Artificial General Intelligence (AGI): You can reason, solve problems, think abstractly, and make decisions.
• Level 3: Artificial Super Intelligence (ASI): It is superior to human capabilities in many areas.

Some Experts | We believe that, given the scope and depth of GPT-4’s capabilities, it can be considered an early but still incomplete version of an Artificial General Intelligence (AGI) system, i.e. Level 2, although this seems debatable. Futurologist and professor Amy Webb, on the other hand, doesn’t expect AGI until the 2040s. Whatever the case, the time to reach a Level 2 system doesn’t seem far away. There is agreement that artificial general intelligence will be followed by artificial superintelligence (level 3). And it is its potential capabilities that are already worrying experts today. AI superintelligence will have a level of power and capabilities never seen before. One hundred years ago, AI pioneer IJ Good predicted an “intelligence explosion” through AI, expected even before it enters stage 3. Machine intelligence will then be significantly greater than of humanity.

The path to get there is already outlined: from level 1 to level 3, we will see the development of intelligent, thinking machines eager to constantly optimize themselves. You will make strategic and operational decisions independently. Imagine a highly connected and AI-controlled smart home. It’s certainly nice if the oven is turned off as soon as the roast threatens to burn, or if the refrigerator keeps track of the food in it and the shopping list is created automatically or sent directly to the relevant stores, which then deliver them automatically.