Applied vs general AI

What we have seen so far with IBM's Watson, ChatGPT, and Google's many AIs, is not true intelligence. All of these examples rely on neural networks - a technique first invented in the 1980s, that has seen a recent renaissance, due to the rise in computing power. But neural networks are not really intelligent. They merely learn to simulate one specific form of intelligence. When you consider the fact, that a NN needs millions, or even billions of examples to learn what a cat looks like, you know it's not like human intelligence. After being trained, a NN reaches a high degree of certainty (something like 95-98 percent), but not entire certainty, in determining if a given image contains a cat or not. A child, on the other hand, merely needs a handful of instances to learn what a cat is - and learns a lot more in the process: how cats behave, what sounds they make, and so on.

Neural-network based chatbots like ChatGPT kind of work like a dog, that has learned a trick - and just like a dog, it does not really understand what it is doing, cannot put it into context, and much less transfer what it learned, to apply it in a different situation or domain. So if NNs aren't really intelligent, then what is? Well, we claim that our approach yields "true" intelligence. True intelligence is not just having one faculty, but a collection of capabilities: recognize patterns, draw inferences, detect analogies, see parallels, form originally new concepts and ideas. askabl does all of this: it can think for itself, critically as well as creatively - just like a human being can. Its knowledge and capabilities will only be limited by the amount of computing power driving it, and its access to information. Another aspect that differentiates askabl from bots like ChatGPT, is that we are not looking to create a chatty AI. We want to provide a more serious tool, that helps to both explain, and drive forward complex issues, like the ones found in science and research.

The core algorithm

The difference between applied and general AI is crucial to the project: the importance of an algorithm, that enables the machine to think at the level like a human being, can hardly be overstated. This algorithm is situated at the very core of our AI. No technical details of how this part works will be disclosed, due to reasons mentioned here. The core intelligence is one homogenous block, which directs and structures the in- and output of a set of neuronal nets, all serving one specific task. So the core collects the bits and pieces, which are its input, and makes sense of them. Surrounding the core, a set of semantically trained neural networks go to work, each of them searching for specific properties, for patterns of meaning they have been trained to detect. When they are done, the core algorithm steps in: collecting and structuring all previously gathered information, and form it into one coherent overall pattern - what we call meaning: the meaning of a thought, the piece of imagination in our heads, when we have an idea, or the process of critical thinking.