I am sure there are sophisticated applications of Artificial Intelligence (AI) in use today. But, SIRI, Echo and Alexa are not what I would consider AI. In fact, I don’t experience AI in my world much at all. They are more a result of good coding, a ubiquitous internet and Internet of Things (IoT) widgets. Sure, put them all together and it looks like the AI we were promised when we watched the Jetsons. They have aspects of AI like knowledge, language processing and in some cases learning and pattern recognition. But these are not AI with reasoning, perception and creativity. Oh, and there has been a raging debate since the 1940’s about the definition of AI.
Look at this high level, functional model and try to see where the AI is:
The modern era of AI started with philosophers, like Hilary Putnam (Born: Jul 31, 1926 Died: Mar 13, 2016), documenting the possibilities of the “philosophy of mind” and then extrapolating those concepts to technology. Our pace of technology evolution has sped up enough so that the philosophers today can see their concepts come true in their lifetime and often even seemingly ideas are simultaneous with technology development and advancement. I was part of an AI team in the mid 1980’s and we deployed successful knowledge bases with interactive capability to replace human decision-making in medical claims adjudication. In this case, before the ubiquitous internet, the design was a database of information, collected from doctors, nurses and health insurance adjusters and a software package written to read and analyze the medical claims details. We called this artificial intelligence because it replaced a team of medical and insurance experts, which saved money, time and the waste of good resources. And, at the time, technical limitation allowed those of us working in AI to dismiss the aspects of AI that were not technically feasible. As I looked for the definition of Artificial Intelligence online today, I was a bit surprised that the definition and discussion is not much different than the 1980’s. AI today is even simplified to dismiss those aspects which remain challenging for technologies to address.
It is clear that sub-sets of AI like voice recognition and visual recognition are the kinds AI that have improved some since the 1980’s. But, would you bet your life on these technologies. We are not there yet.
At COLOTRAQ, we are addressing the needs of clients who demand data centers and cloud services supporting the latest AI generation. Our data center designs often are geared toward micro/distributed data center designs to optimize edge computing and network response. This is all part of the ”magic” under the covers making this new run at AI possible. We are also seeing AI at play to replace the data center operations staff, at least the night shift. And, we are hopeful that the AI in data center cybersecurity will stay ahead of the AI being developed by hackers. But, we humans will still need to look under the covers, understand how it all works, and test to be sure that we are optimized, safe and secure.
This article was originally published at COLOTRAQ blog.