Can Artificial Intelligence
Replace Humans?

Can Artificial Intelligence Replace Humans?
What Are The Risks Of AI?
Will the escalating growth of artificial general intelligence (AGI) lead to human extinction? The gorilla survives only because of human goodwill. The superiority of human intelligence currently props up mankind above other species. Computers today greatly surpass human skills. Each generation of computers improves on the previous one. Many fear that a super intelligent AGI could tend towards the so called instrumental convergence.

Their unbounded goals of self preservation and resource acquisition could cause an AGI to be lethal to humanity. An AGI commissioned with a limitless goal of solving a mathematical problem could seek to marshal its computing power by turning the entire Earth into a computer. Instrumental convergence would cause it to resist attempts to shut it off or to change its goals. Stephen Hawking, Bill Gates and Elon Musk have expressed concerns about AGI.

  • What Are The Risks Of AI?
  • What Is The Starting Point Of This Article?
  • How Are Human Limbs Controlled?
  • What Is An Artificial Intelligence Program?
  • What Are The Capabilities Of AI?
  • How Can Humanity Control AI?
  • When Do Humans Cede Control To AI?
  • What Are The Ethical Considerations For AI Controls?
  • How Can AI Control The World?
  • Is Humanity Stupid?

Over 2 Million Page Views - How Did It Happen?
I am not a physician, but an engineer.  Way back in 1989, I noticed the unique speed of an AI algorithm for disease recognition and pieced together the implications of that phenomenon.  The same algorithm could  be applied by the nervous system for the management of its stored knowledge!  The axon hillocks of neurons could store and retrieve memories of trillions of links between symptoms and known diseases.  Neurons respond to received impulses in 5 milliseconds.  The axon hillocks of neurons could use an elimination algorithm to eliminate (inhibit) unrelated relationships.  Since its symptoms were not linked in memory to any other known disease, a doctor could locate one disease in milliseconds.  

Functional regions could use precise combinatorial codes in feedback and feed forward links with incalculable eloquence.   The axon hillocks of the neurons could store and retrieve memories of ancient jungle encounters and herd struggles. They could carry memories of the indications of danger, of strategies for survival and of motor control responses.   One region could link sensory signals to potential trouble from people and things. That could cause other neural organs to trigger anger, or fear. In turn, these emotion signals could make a person flee, or fight.  An insult could trigger stress in a split second.  On the other hand, the self awareness tool could inhibit stress and still the mind.

Not metaphysical theories, but millisecond pattern recognition could explain human brilliance and folly. It could also offer the means for its control. Several years after my fortunate insight, a Nobel Prize acknowledged the astonishing power of combinatorial pattern recognition.  Over 30 years, I have assembled much evidence supporting neural pattern recognition and published three books about artificial intelligence and the nervous system. 

Can Artificial Intelligence Replace Humans?
How Are Human Limbs Controlled?
The human mind is the most powerful known intelligence in the world. Yet, David destroyed Goliath with a well directed pebble that disrupted the prefrontal regions of the giant. Those regions function through a hierarchy of self governing sensing, feedback and control systems. They operate in a manner similar to the governing processes of a country. At the lowest levels, individual actors make their own decisions. 

Higher level decisions are made by the communities, by the state governments and finally, by the central government. Experiments revealed that if the regions above the thalamus were cut off, the intelligences below could still manage homeostasis - “feeding, drinking, apparent satiation and copulatory responses in a broad range of adverse conditions." But, shorn of purposive guidance from their central command, the mighty limbs of Goliath could merely twitch in convulsion.  An intelligence can also function only when it is goal driven.

Can Artificial Intelligence Replace Humans?
What Is An Artificial Intelligence Program?
An AI program uses knowledge to achieve a specified goal, while meeting a sequence of conditions. An AI program is not intelligent if it lacks purpose. It is an axiom that purposive choices of actions lead to goals. Purpose driven cellular life survived even the demise of the dinosaurs. Purposive human choices have created untold health, wealth and prosperity. Computer programs use algorithms to achieve the goal of solving a problem through a sequence of unambiguous instructions. Such programs may be written by humans or by AI programs. It is unacceptable for a program to produce random, or stupid results. It must meet well defined goals and limitations.

Can Artificial Intelligence Replace Humans?
What Are The Capabilities Of AI?
There are powerful AI programs that improve the quality of government services. They study data and detect patterns and generate rules in situations where there are no rules for behavior. They identify hitherto unknown patterns and relationships. They detect anomalous or unexpected behaviors. They predict likely outcomes or actions based on identified patterns. Their performances improve with more data. They can analyze your vitals, correlate them with your medical and travel histories, and predict diseases, or likely treatments with accuracy. But, they all perform specialized tasks. Shorn of higher level guidance, they are not true human adversaries. They are currently the limbs of the human intelligence.

Can Artificial Intelligence Replace Humans?
How Can Humanity Control AI?
Humanity can be destroyed only through physical actions. AlphaGo can defeat a professional Go player, but cannot lift a feather. For large scale control of physical equipment, AI requires products made from steel, concrete, silicon,  open cast mines and large capital investments and statutory approvals. They require the approval of not individuals, but of large populations. Even oil pipelines are halted if they offend public concerns. The modern population is articulate and perennially alert to the dangers posed by AI. Such a society is more than likely to ban public AI uses, which could be considered to be even remotely dangerous.

Can Artificial Intelligence Replace Humans?
When Do Humans Cede Control To AI?
AI self driving intelligences form one of many limbs of large scale AI control. Such AI controls a limited number of automotive actions. They can start, stop, go forward, go backward, turn, speed up and slow down vehicles. Across billions of miles of operation and many generations of programming, such an AI must keep humans safe. Safety is mandatory in varying road and weather conditions, presence and behavior of other vehicles, people and animals and the rules of the road.

For the physical control of social life, each limb of AI must prove the safety for humans. Such safety is not the will of the AI program, but its essential inbuilt structure. Even for military assets, humans ultimately controls the goals. Despite dire predictions, humanity has not so far exhibited the unrestrained suicidal inclination needed to hand over the control of nuclear bombs to AI programs.

Can Artificial Intelligence Replace Humans?
What Are The Ethical Considerations For AI Controls?
The AGI that controls law and order tools such as swarms of robotic police, drones, crowd control vehicles, prison facilities, arrest and release functions has the power to destroy humans. Such AGIs will face overwhelming ethical problems. A typical one concerns the choice of a switch that decides whether a trolley kills one person on one track, or five people on an alternate track. An AGI could, of course, have access the knowledge of a safe choice of stopping the trolley.

When an AI program was ordered to make a 4 legged robot to move without its feet touching the ground, it commanded the robot to roll over and move using its elbows. But, Hitler, a human, decided to kill six million jews for a presumed better life for billions of succeeding human generations. Human indignation against police atrocities is only a post disaster tool. Whether AGIs can safely make these decisions may continue to be debated for generations. A kill switch for the AGI at this level could be the only pebble that can kill the AI Goliath.

Can Artificial Intelligence Replace Humans?
How Can AI Control The World?
Each time they are proved safe for humans, AGIs have an inescapable superiority. If all the limbs of such a structure have been proved safe for humans, AI programs could take over all functions of a government, such as law and order, finance, defense, health, education and economic production through a hierarchy of country, city, local and village level systems. When every limb of an AGI government has been proved safe, a central AGI, (or an AGI committee) could exercise control at the top through the cloud from multiple parallel, protected hardware locations around the world. Such a central AGI government which controls physical assets could have the power to destroy humanity.

Can Artificial Intelligence Replace Humans?
Is Humanity Stupid?
Technological innovation has perennially enabled ordinary people to enjoy extraordinary improvements in standards of living. Despite a totalitarian regime, China has also advanced rapidly.  But, society, as a whole, constitutes a super intelligence, with an incredibly painful sensitivity to the dangers of AGI. They also have the legislative, capital and investment power to hold AGIs in check. AGI is not a multi-headed hydra that can just snuff humanity out. The majority of the limbs of even such an AGI are harmless. All their program structures demand goals and restrictions.

But, such AGIs do require a central control head.  But, its lethal aspects should always be under human control. It is difficult to believe that a future human generation will be stupid enough to omit a simple enough kill switch for the hazardous parts of AGI that can harm humans. Despite doomsday forecasts, humanity has had its thumb on several AI controlled nuclear buttons for three quarters of a century and still not committed suicide. It also does not appear as if, in the foreseeable future, human society wii be dumb enough to allow itself to be replaced by Artificial Intelligence. But, you never can tell.

This page was last updated on 31-May-2020.

Jordan Peterson - Happiness
Can Artificial Intelligence Replace Humans?

JUST THINK...   What happens when you begin to talk? Your nervous system has picked an emotion. It has articulated an idea around it, chosen apt words from a vocabulary of thousands of words, arranged them in lexical and grammatical order and adjusted the pitch of your voice. Before you speak you've no consciousness of the words you will use. Who's actually in charge?  This question leads to the question "What is consciousness itself?" Is consciousness a spirit living in a human body?  Is it a mystical life form that emerges from the nervous system?  This is the hard problem of consciousness.