Definition: AI complexity

With all the automation that has taken place in the last 75+ years, the complexity within the digital world humanity has created is monumental. The sum total of software modules communicating with other software modules running 24/7 is staggering. And all that was before AI.

An Irony of Complexity
There is no doubt about the huge amount of complexity AI is adding to the digital world at every level as it simultaneously attempts to hide that very same complexity from users. When an AI chatbot delivers an erroneous result, the AI programmer may not always be able to figure out why. Although AI is increasingly used to debug AI, there are times when it cannot. Unlike regular data processing applications, AI is a totally different paradigm. See AI debugging.

In addition, due to its unpredictable and sometimes massive power requirements, AI is adding an enormous burden to the U.S. electrical grid. AI solutions are attempting to address this complex problem with clever grid management. See AI datacenter and Tesla Energy.

What started in the 1940s when only a few dozen computers were operational worldwide has migrated to more than a trillion computer-controlled devices in every nook and cranny on and "off" the planet (see space datacenters).

In Just Two Years

Everyone is deluged with complexity. This is the number of files scanned in a Mac desktop computer used for two years by one person. If this is the number of software components for a single user in two years, imagine the total number of digital files in a company with ten thousand employees after a few decades. AI only adds to this jungle because its inherent architecture is a maze of complexity. See neural network.