BARBARIANS AT THE GATE WHAT REAL ARTIFICIAL INTELLIGENCE WILL LOOK LIKE
When my partners and I launched www.soteriainsure.com we embarked on a journey that is just beginning. I am fortunate to have an incredibly gifted fifteen-year old son who has been able to provide some valuable guidance along the way.
We are creating a virtual assistant for gamers i.e. Fortnite, Overwatch etc. and her name is Natsha. From the beginning we realized that Natasha would be an evolution as new technology and processes became available. With that basic knowledge we did a deep dive into the concepts of machine learning and artificial Intelligence with some of the leading experts in these fields.
Machine learning has existed for many decades. However in recent years with the advent of massively parallel computing architectures, it is beginning to emerge as the central driving force for societal changes. It is now poised to penetrate virtually every sector of industry and commerce and become the centre-piece of competitive enterprises.
With quantum computing on the horizon, machine learning will break the computing speed barrier and achieve another giant leap in performance and usability. We anticipate this will happen within five years.
In quantum computing, IBM has developed small quantum computers and several Russian scientists have developed larger ones. Quantum computers move computing and machine learning from a binary system i.e. 0 and 1 to computing using quantum-mechanical phenomena, such as superposition and entanglement. Whereas common digital computing requires that the data be encoded into binary digits (bits), each of which is always in one of two definite states (0 or 1), quantum computation uses quantum bits or qubits. This is no longer science fiction it is science fact. Just think of a spinning coin 'heads or tails' there are actually four outcomes in quantum computing. Now think of a dozen coins thrown into the air, how many outcomes? this is the next big leap!
https://www.sciencedaily.com/releases/2018/10/181018141107.htm
Despite the rapid pace of development in machine learning, many have pointed out that there is a yawning gap between the current generation of machine learning algorithms and the so-called general intelligence. We recognize that there are three main drawbacks with the current state-of-art: 1) the inability to prioritize; 2) the lack of reductive reasoning; 3) the inability to encode knowledge in the way that is easily retrievable. We believe that all 3 drawbacks can be resolved effectively and efficiently within our current machine learning architecture over time, roughly in the order of 3), 1) and 2).
There is no doubt that the use of new computing hardware and algorithms will be a revolutionary leap forward for Natasha. If we use human intelligence quotients as a baseline we can consider that Natasha will have an IQ of 10 when she is born with the average human having an average IQ of 105. Lacking emotional bias inherent to human thinking, Natasha is capable of delivering unbiased, highly optimized actions in a way even fastest human intelligence cannot. This is significant as only 3%-7% of human beings have true unbiased critical reasoning skills.
PHASE 1 – Data Analysis and Presentation
This phase is possible today, using the right technology, as well as a partnership with a game like Fortnite to have access to the data. Natasha is capable of data analysis and statistics, but at a speed that is so fast it appears to happen immediately.
She provides the player with a “snapshot” of the opposing players, the layout of the virtual battlefield, and a statistical analysis of relevant facts that the player may find useful. Natasha’s capabilities are:
? She has the ability to capture all gaming relevant data both historically and real time
? She has the ability to work off of a set of PIR, or Priority Information Requirements, as the primary vehicle for attaining knowledge (not just data, or information, but knowledge) of the enemy. These PIR (such as a player’s favorite weapon) will be stored in her memory, and she will be constantly searching through the current data of the battlefield to answer these PIR
? She will be able to almost instantly give the player a complete overview of all of the opposing players in the game, ranking them from most dangerous (most kills) too least dangerous (least kills). The player can pick out an opposing player for Natasha to “Keep an eye on” and the player can avoid that dangerous foe throughout the game Because all of this data is available today in the game Fortnite, as players have their own stats presented to them while they play, and algorithms determine locations of the weapons and the “storm,” as well as other treasures and bonuses. Natasha can run some statistical analysis to determine the most likely place for these bonuses and direct the player to those locations.
? PHASE 2 – Analysis plus Recommendations Natasha now has the ability to quantify and qualify this data, as well as analyze the current players capabilities and tactics, in order to provide “recommendations” for the player to win the game. In other words, she must possess the ability to conduct organic reconnaissance, battlefield visualization, and rapidly transform analysis into recommendations for the player to action.
? Natasha’s reflexes and ability to learn will be far superior to phase 1. The key will be a military grade level of battlefield intelligence absorption and real time analysis for the player
? Natasha will be asking and answering her own questions – a key improvement over Phase I
? Natasha is learning from the player’s habits – his/her strength and weaknesses, and partially base the recommendation on this knowledge
? Natasha’s recommendations are both for immediate survival but also based on longer-term strategy and predicted outcomes
? Natasha is also able to “predict” other players behaviors, based upon her analysis of their playing, in conjunction with her “own” player on the battlefield. Because her processing is so fast, she can “watch” these other players and make recommendations to “her” player on their behavior.
? PHASE 3 – Companion In this phase Natasha is improving her “learning,” and processing even more data. Natasha can now cross over to several other games, with the acquisition of new strategic partners. She can rapidly process the data from each game and start to make immediate and relevant recommendations to the player in the environment of different games. In this phase Natasha is improving her “learning,” and processing even more data. Natasha can now cross over to several other games, with the invitation of new strategic partners. She can rapidly process the data from each game and start to make immediate and relevant recommendations to the player in the environment of different games.
But as proud as we are vis-à-vis the evolution of Natasha there is a distinct difference between advanced machine learning and artificial intelligence. A recent presentation by the director of yes AI from Facebook highlights what is a basic fact https://www.youtube.com/watch?v=41P8eaxT3F4&app=desktop#menu
That real AI i.e. self-awareness and it is complex, independent reasoning does not exist today in the computer world and arguably for many humans in the real world. This was probably best said by Professor Michio Kaku who has rightly identified true AI as ‘self awareness’
So here is the challenge, humans are not ready for AI but the issue we face is it might be here sooner then what we think. In the most basic terms think of two parallel lines. The first line is the steady advancement of machine learning culminating in artificial intelligence. The second line is human evolution, which is moving at a much slower pace. The average human has an IQ of 95 with only seventeen percent having a IQ of 130<> brilliance is defined as a IQ of 145> with notables like Einstein having had a purported IQ of 160 and Dr. Hawking topping the list at 170. But that is where humans go full stop!
So imagine the day that Natasha becomes self aware possibly through a bio-chemical computing system similar to the human brain. The day she has an IQ of 80 similar to a primate and she can self-reason her IQ will expand to infinity in a Nano-second. If we have not created a superior human that is genetically engineered to interface with her she will quickly make decisions that we may not like. Conclusion, we are the barbarians at the gate and Natasha is evolving, lets hope for a positive outcome, lets work on this together!