Here is a description of what constitutes a Good AI and what forms part of the Bad AI. There has been tremendous progress in AI, irrespective of the fact that it is a good AI or it is a potentially harmful AI. Let us discuss what is in good AI and what is in bad AI. All these are motivated by current discussions and questions about AI development without checks on the dangers of AI.
There is always a danger in every new field, for example, diamonds, which many people like so much, diamond mining too need the passage of dark mines with no light, which can be dangerous. So, how does the danger of mining not stop us from picking out the best diamonds? This is called by right caution, guidelines, prohibition, and following the rules. If all that is followed and adhered to then the right mining can be performed.
In the same way, AI mines are getting darker with least lights to see, what was on the top surface has been used and extracted. When you go inside the mines of the science of AI, then there is darkness, no light and uncertainty. Nobody knows what is coming out of it mining the AI deep mines. Let’s learn from Earth Miners what is safe versus what needs to be ignored in earth mining.
The good stuff that comes out of AI is like those AI that help businesses, technologies, medicines, crime detection, many other subjects, and many other fields. These need predefined output guidelines, and these AI can be inculcated in AI software which can be used without doubts of wrong outputs being generated. These include but are not restricted to sales leads, water purification, washing machines, auto suggestions in text, and so on. These AI are being used and can be viewed as the beneficial or good AI, as all they are doing is not causing anyone harm and are just helping.
The other side is bad AI. Imagine this as a week part of earth mining, where things can be dangerous if not taken care of and there are parts of Earth mining where one should not go to. In the same way in AI development there are things that one should not try, such as making LLM learn a wrong context, or making harmful entries in AI and LLMs in particular, allowing young children to use AI is also wrong, as they are not adults they can’t decide what to do with AI, so why they haste of teaching them AI in schools, when they are not adults. Many more AI apps are in bad AI, such as those that make beautification of a selfie, AI games that involve playing in with AI agents, to even adults AI apps that fakes people profiles with audio and many more.
On one side people see AI as money-making entities, and when it does wrong, for example becoming an addiction or effecting self-harm, in that case, one realizes that it’s too late to keep the safe mine with lights on. So don’t run for AI development, make plans, get them sanctioned and then go in for AI development. Keep kids away from both AI development and AI user applications, which are a real threat to child development.