How Likely Is the Coming of Skynet?

How likely is the coming of Skynet (or other destructive artificial intelligences)?

The coming of Skynet as such is entirely possible and yet still very far away from a realistic problem. This can mostly be attributed to the fact that "modern" AI is still incredibly far away from functioning at the level required for a "Skynet"- type problem to occur. Our AI has improved hundred if not thousand folds in the last decade; but even then it is still insignificant in comparison to the most simple AI that could rival a Skynet AI as thought of in the Terminator Franchise. Author of the famous book "SKYNET TODAY?"(2014), Albert Zweistein, explicitly delved into the simpler mechanisms behind creating such AI. He looked at problems such as gravity and power consumption being the limiting factors for growth in AI to the levels that would allow a Skynet to control machines such as those depicted in dystopian movies. Simply put, his finding conclude that Gravity would impose huge restriction on the size of machines (robots as we know them) that could realistically still move. The book takes a very specific look at robots of the magnitude shown in the movie "Pacific Rim" (2013). Being such an incredible physicist, he worked out that gravity would simply not even allow the robots to move as fluidly as shown, let alone swing a sword. The sword, as he elegantly puts it, would "flop in the wind" (Zweistein, 1999, p.67). Looking at this in detail shows that it is not simply a matter of the robot being unable to lift such a mass, here a sword, it is also a matter of physics where such objects simply cannot be lifted against gravity for the intended purpose. Zweistein's outlook on the matter is simple, robots being controlled by a Skynet-type AI simply cannot exist because of gravity alone and if one were to somehow defy the laws of gravity then the electricity consumption would rival the daily usage of electricity on the entire planet "According to an IEA estimate, we humans produced and used  5.67 × 1020 joules of energy in 2013, equivalent to about 18.0 Terawatts which is just short of what it takes to power such a theoretical robot and it's movement for a period of 15minutes" (Zweistein, 2015).

This energy consumption estimate is the equivalent of 5 billion barrels of oil. Therefore, we as a race simply cannot even begin to power robots of this magnitude if we had the resources to build them or control them which leads to the conclusion that Zweistein alludes to in his book titled "Delving into the abyss of robotics or why we will never make it" (2015), this being that even if we managed to develop the necessary AI, it would not have a home to be kept and therefore be superfluous to our society.

Another important concept that Zweistein brings attention to, is that much like the current day Pharma-industry, if a process or an invention were to generate no financial benefit then it is very unlikely to ever see the light of day. He does admit that blue sky research into this matter is possible but even then, the only government with the required knowledge, financial backing and qualified staff would be that of the Democratic People's Republic of Korea (commonly known as North Korea).

Assuming that the intentions to create AI and Robots that can create a use for this are followed through, the aforementioned problems related to the physics behind the robots would leave them at an impasse. Assuming the AI is made purely to infiltrate out current systems then it is simply not possible because even today in our current technologically advanced state, we still run on archaic technology for the most part. The binary system is ages old and yet still in use, tertiary systems simply cannot even begin to rival the binary systems that are in place and will likely not do this for a very long time. Whilst there has been some research into fluid based tertiary systems (Ymous et al, 2010), these have proven to be unreliable and require infinitely more upkeep than our current technology does. Additionally the space constraints cause the tertiary system to fall wayside, requiring almost 100 times more space than a regular current generation AI would. Ymous et al look extensively into the problems behind the current generation AI and the Skynet dilemma in their book titled "Why we cannot have our cake and eat it too". The conclusion found resembles a very simplistic approach to the largest issue that is commonly posed, "how would we deal with a skynet occurance?". The solution that Anon Ymous offers in his book is found in a chapter that he aptly titles "unplug your toaster". The solution today and for the foreseeable future is to just unplug your electronic device in order to undo a rogue AI's ability to control our world. Current generation computational devices simply do not hold the processing power nor the battery life to do any significant harm let alone any insignificant harm. Ymous makes reference to the largest terror attack by a rogue AI in the current day that he could foresee being that "all our handheld devices refuse to connect to WiFi thus causing us to incur large 4g charges with our carrier, that is about as bad as it could get" (Ymous et al, 2010).

...