Artificial intelligence” as we call it today should be termed as “Perceived Intelligence”/“Pseudo Intelligence”. Why ?
From #deeplearning to #machinelearning to #naturallanguageprocessing , to #ai as we know today, the fundamentals are still in data. We feed data into the systems, get them to “learn” and then mimic what a set (1-xMillion) of humans could possibly do.
The gap in all of these AI systems today are that they skip the required essential intelligence/reasoning/behaviour/awareness capabilities . This by nature is not possible within the building blocks of AI as we know today. For example, the layers within deep learning libraries such as Tensor/Keras/GPT3/OPT tries to mimic/perceive neurons as the building block for a neural-network, which is fundamentally wrong when compared to a biological neuron which is more aware of itself.
It is naïve to believe that the building block for an AI system would be software driven as the software is still limited by the hardware it runs on. Scaling still depends on more compute which is managed externally and further more, these AI systems of today are not designed to adapt to its environment based on its observed limitations (awareness). Think of manifesting hardware for software to execute (a neuron needs an environment to grow which too must be dynamic/evolving).
The building blocks for an AI system would probably be a chemical-electrical entity that can manifest its wares (soft/hard) as required; mimicking a biological neuron/dna (?). An AI system should be one where intelligence/thinking/reasoning/awareness is contained within it by nature/design. Would believe these systems would not have a classification of hardware/software as they would/should work as a single “ware”
If IQ is defined for humans, we should be able to measure a similar AIQ too for these systems, if not IQ itself. The system should be self-aware, know itself and its constraints (environment/knowledge/abilities) too.
Until we get to that point, we should change the general usage of artificial-intelligence to perceived-intelligence or pseudo-intelligence. This is important to make sure we can make a clear distinction in communication on any matters related to the *real* artificial-intelligence (perhaps not in this lifetime).
#artificialintelligence #datascience #datascientists #innovation #tech #python #coding #programming #perceivedintelligence #ai
