The actual status in AI development

There is neither a valid universal model for creating a general purpose AI with a body,
nor a corresponding business model to commercialize it, because it does not exist yet.

Assumptions and constraints

A nervous system, either biological or artificial, always adapts to several major factors like:
the pre-defined body, the environment, the interactions, the accumulated experience, etc.
Even if 2 bodies have the same components and structure, say biological twins or identical robots,
each nervous system is unique, due to the accumulated experience during its interaction in the environment.

ML applications - specialized for only one purpose

Most conventional neural nets are used as a gigantic database with millions of items,
all of them containing only slightly different views of the same object type, e.g. a cat.
The well known ML, Machine Learning, by using millions of data records is not AI,
instead it is a just statistical method, which enables a guessing of a certain pattern.

General purpose AI using biomimetic model

The first hurdle for our company is to explain to others, that our AI approach is valid,
while following this path inevitable leads to achieving the goals of a general purpose AI.
Taking the biological life as a template for developing a new technology is always a valid path,
therefore new sciences are born, like bionics, meaning biologically inspired engineering.

The second step is to prove that parallel processing is the only viable method to be used in AI.
Conventional computing has proven its advantages in many areas, where response time does not matter.
Biological intelligence requires another way, namely parallel processing, enabling rapid response times.
Every neuron within every neural net is processed simultaneously, all of them reacting within only 1 ms.

Third, we have to create first a proof of concept, which is our operating system,
which should be hard-wired and be able to process all kinds of neural nets.
It will represent the basic functionality of our biological human brain,
which consists of only neural nets and nothing else, except for power supply.

Fourth, afterwards create a prototype, which will showcase AI applications.
This is really sensitive, because there are many brains  with different grade of complexity,
each representing an own level of evolution, from insect via mammal to primate and human.
But all have a common feature, namely an embedded self-learning mechanism to enable adaption.

For this purpose, we have to invest a lot of time, patience and many resources,
which includes the decision, which body hardware is most suitable,
not to forget the end result, by showing our results in a videoclip,
so that other companies can see the abilities and skills of the AI.

Nevertheless, the legal implications of using AI in the real world are unknown,
since there are no laws for this purpose, we barely begin to formulate them,
even if this creates another ethical problem, how to define intelligence at all,
e.g. the stakes for safe autonomous cars are way too high, due to insurance reasons.

There is another very big obstacle, namely the public opinion.
Most people do not even want AI in their life, for different reasons.
Ironically, some CEOs of leading technology companies which develop own AIs,
are warning that someday in the future an AI would destroy the entire humanity.