GPT-3 is an example of multi-model AI which is setting the standards of how complex connected AI models can deliver beyond average, human-level intelligence. This is where we need to stop or control general intelligence from talking to other general intelligence, otherwise, they will eventually enable ‘super intelligence’ scenarios. Then, superintelligence talking to superintelligence could create something we don’t understand, and we may lose control. This is where the possibility of AI creating a new form of intelligence may be viable. It’s catalyzed by the rise of GPU today, but it will be Quantum computing soon: at some point, we will experience a new level of technology, one that we don’t know about yet, and we may realise that we are dealing with the natural evolution of advanced technology systems.

Advancements in AI could result in an Intelligence offset that we don't understand, and if this is applied to autonomous Machine Learning, it will result in a paradigm shift in our society. Like Genetically modified foods, Artificial Intelligence may enable ‘digitally modified citizens and 'Intelligently-controlled-societies', and once this happens, there may be unthinkable hidden possibilities and risks that we cannot even fathom. The danger is we facilitate an uncontrolled and unregulated world that doesn't recollect how to rationalise i.e. apply Human Intelligence.