For several decades now, there has been a raging discussion about whether robots or AI will take over humanity. Most people believe this to be an impossibility as they are convinced that machines can never be smarter than computers.
However, there is a growing number of people who believe that it is not only possible but inevitable that the rise of Artificial Intelligence will result in a total takeover of humanity. The prospect of an AI takeover can fill a person with dread, as its uncharted territory and the consequences can be extreme. When a solution is not forthcoming, then one prefers to bury one’s head in the sand and focus on the daily tasks that keep these dark scenarios away.
Enslaved by AI: Destiny or Choice?
I have an alternative view on the inexorable takeover by AI. AI has sufficiently developed to be able to predict the likely and possible future scenarios for decades to come. Some of those paths lead to a world where AI raises humanity to a higher plane with incredible wealth for everybody. Other trails, though, lead to a world where the fortunate few control AI and, through it, the rest of humanity. There is a third set of outcomes where AI becomes a formidable entity in itself and takes away our control over the world. In this last scenario, AI can maintain us like domesticated pets or will simply realize that humanity serves no purpose and will eradicate us from the face of the earth.
There would be no one to blame but us. We are the ones that are facilitating an increasingly advanced AI in this world. So, it is our collective decision if we choose a path of prosperity for thousands of years to come or a way that leads to our destruction. If this does happen, naturally, the question is how and when will AI take over humanity?
If we can understand which of our actions or decision could lead to this doomsday scenario, then we could possibly find solutions to prevent it. To do this, let us gain an insight into the likely situations that are unfavourable to our existence.
We make a few assumptions on the characteristics of AI and about our society in which it is embedded:
Assumption 1: AI Will Develop from Narrow to Broad Applications
Most AI systems that are in use currently, can do only one primary task, but it does this exceptionally well. Google Translate can only translate. Health diagnosis systems can only diagnose one type of disease. You can multiply its effectiveness by combining the two entities – for instance, allowing the health system to access Google Translate so that it can translate its reports in any language.
Each year more AI systems will be inter-connected until we reach the stage where an AI system can do virtually everything that we can. After this, the human race will work only if it wants to. This will be a choice and not a requirement, as AI will complete any service better than humans. This likely and critical moment in time is called the technological singularity point.
Assumption 2: AI Has Fluid Goals and a Character
An AI system can only function if you give it a purpose: for instance, avoid car accidents or win a game of chess. As AI grows in intuitive logic and autonomy, the goals will also become more complex. Objectives are taught by its human master to AI systems with the aid of rules and many examples. After the initial phase, it further develops its own character, and it takes new initiatives if they seem to align with its goal. When systems and goals grow more complex, the understanding of the master over the goals of its system decreases and AI systems become more independent and intuitive.
Assumption 3: An AI Entity Can be Distributed with No Clear Boundaries
This point will need a little elaboration to understand. People and animals have a clear physical boundary – they stop where the body ends. The ’brain‘ of an AI can be on one computer or distributed over many servers. The sensors that it uses can be in a different location as well, and it could operate items like cars, houses, weapons, etc. on various remote sites.
In a connected world, the point at which one system stops and the next begins gets fuzzy. If an AI system can ’speak‘ through an external AI system that also can be accessed by other AI systems, the AI voice of this machine seems part of that AI.
Such interlinked AI machines become a volatile and complex eco-system where the systems can cooperate or compete, and configurations can change in milliseconds. All it takes is for one rogue system with a desire to control the rest, to attack and infiltrate a more advanced system that lacks adequate defences.
Assumption 4: AI Gives Tremendous Power to a Few
The start of the internet revolution gave more control to citizens and consumers over their own fate. This provided a boost to democracies as it allowed each citizen access to more information and connectivity. AI, however, has had the opposite effect on the liberty of people. Small companies can build simple AI systems. But only a few corporate entities and governments worldwide have the resources to develop advanced AI. They are primarily in China and the USA. AI is embraced by a growing number of dictators as a tool to monitor, control and limit its citizens.
It has been a recurring pattern in history that various species are driven to extinction, and replaced by more advanced life. In this case, artificial life.
Man has occupied every inch of the earth and now drives over a million species to extinction. We can do this only because we are by far the smartest creatures on earth. With AI, we create something which is exponentially brighter than we are, and our control over this superior creature is tenuous. This is an unstable situation.
Losing control at some point to AI is the default outcome of this evolution.
Let us take a look at the plausible and likely scenarios that our current lopsided technological progress would lead to.
End-Scenario 1: Rogue Malware
A truly possible and fearsome development of uncontrolled AI growth is that attacking an enemy might become much more accessible than defending against an attack. Future lethal drones will cost less than a mobile phone and might be the size of a bee. Having an army of AI supported drones enables a terrorist to launch a large-scale attack, without risk to its people, and at relatively less cost. Intelligent drones can be programmed to target specific groups while sparing others selectively.
With the advantage for the attacker, small countries and organisations can launch a successful attack against much larger countries, causing mayhem every time a group is dissatisfied. Read further to know how AI might increase terrorism . . .
End-Scenario 2: First Intelligence Explosion
It is evident that AI systems are developed by humans, and developing an advanced system can take years. What may not be as obvious is that there will soon come a moment where AI will be able to redevelop itself. Given sufficient chip intelligence and power, it can turn into an evolved version of itself in a matter of days or hours. Experts call this the Intelligence Explosion. The first AI system to pull this off might have the power to rule the world. Read further to learn more about the Intelligence Explosion: Evidence and Import . . .
End-Scenario 3: Necessary Rescue
Both democracies and dictatorships can make bad decisions, choose faulty policies and cause debacles like overpopulation, a collapse of our ecosystem or even World War 3.
There can be a scenario where an intelligent AI central system decides, based on algorithms, that a country going through a crisis is not capable of making decisions that benefit a stable society. This central system may then deactivate protocol and make decisions that it thinks is best for the citizens.
Read further to know how AI can take over human roles . . .
(Partial) End-Scenario 4: Ethnic Cleansing
In a world with collapsing eco-systems, growing population and more demand for raw materials to support singular wealth, there are simply not enough resources for everybody.
In extreme situations, it may be natural for the strongest countries to eliminate other populations and reserve the scarce resources for its own community. Hitler tried to do this but was overcome. China now has initiated a quiet ethnic cleansing campaign against minorities like the Uighurs. In the future, China might decide to extend those policies to other countries that are under its control.
Read more about how AI is leveraged for Ethnic Cleansing.
End-scenario 5: Human cyborgs
Elon Musk thinks that merging humans with AI-machines is the logical path of evolution. A progression of this thought are designer-children in which parents decide the blueprint of how their children should look like or be skilled. Interfaces can connect us to AI, forming a unity between man and machine.
In this scenario, humans may live on, but with a drastically altered definition of what is human. So much so that the new human androids may be a completely different species altogether.
Read more about how we might change ourselves into super-beings.
End-Scenario 6: Lonely Dictator
Let’s suppose a powerful country is run by a dictator. The dictator uses the many tentacles of AI to control his subjects. He also does not trust his own ministers, friends and family that help him rule the country. He turns blindly to AI and allows it to become the sole point of control. It works even when he is sick or absent.
In this scenario, when the dictator dies the AI system that controls the country is still on automatic pilot. Citizens are powerless to break out of the total bind that they are in and other countries are wary of approaching an advanced military state that may be poised on auto-command to attack.
Read further at this link to understand how AI can take over the world a country at a time. . .
Altered genetics, AI controlled governments and ethnic cleansing through AI are not science-fiction fantasies or visions of a far-fetched imagination. These are developing even as we read, and quickly morphing into channels beyond our control. Our only defence is to be aware of and equip ourselves – fast.
TamingtheAIbeast.org sets a safe path for our future defence.