Does AI pose an existential threat?


We are going to have tsunami of Artificial Intelligence in every aspect of our lives. Already intense use of AI applications are in ecommerce, search engines, automated logistics, hotel/airline ticketing, stock market, computer games, mobile, quantitative trading, robotics, healthcare, handwriting/speech/face/emotion recognitions, self-driving car, surveillance, space, military etc.

AI has taken over human roles in many walks of our life. Experts estimate that within a decade, net 47% low and high-end job loss will be due to AI. Stock markets run by AI and human beings plays secondary role. The pricing of hotel room and airplane tickets is managed by backend AI.

Some experts feel that AI pose greater existential risks for human race than climate change, geo-engineering, global pandemic, molecular nanotechnology, nuclear radiation, synthetic biology and human modification.

Experts feel that the automation progress has been exponential in two decades, AI will surpass human capacities in all respect. Progress from Human level to Super level will be much faster.

Regulators should look at biotechnology and nuclear guidelines that avoided the dangerous mutation and radiation. AI researches can embed ethical control as in the life science, medicine and genetic with Hippocratic Oath.


AI is real and inevitable: Let us decode Artificial Intelligence. AI is an automation process which has evolved since industrial revolution. Machines can analyse and digest data million times faster than human. AI does not replicate the physical brain but its functions; similarly as airplane imitate flight task, rather the bird.

AI researchers have achieved critical milestones during the last few decades. These milestones are the Decision theory (study of the reasoning and choice), Heuristic search (faster good choices to eliminate bad paths early), Decision tree (to identify a strategy most likely to reach a goal), Evolutionary algorithm (evolutionary computation inspired by biology), Temporal Difference learning (prediction-based machine learning method), Big Data (large and complex data needing new processing software), Variational auto-encoders (artificial neural network used for unsupervised learning), Convolutional Neural network (connectivity pattern of neurons inspired by the animal visual cortex), Deep learning (class of machine learning algorithm), Neural Turing machine (extension of neural networks) and many more.

Experts feel that the automation progress has been exponential in two decades, AI will surpass human capacities in all respect. Progress from Human level to Super level will be much faster.

We have momentous phases for examples IBM AI has beaten world chess champion decades back, Alpha Go defeated world champion, Google AI diagnosed tumors better than radiologist did and IBM Watson outclassed physicians.

Ultimately, artificial general intelligence will make technologies disappear and seamlessly work in the background rather the alerts driven nags like email, tablet, mobile etc.

Radical Progress: True Artificial Intelligence is when computer writes algorithm itself without human intervention. Algorithm that learn, update, procure resources over the Internet and improves itself making AI many times smarter than human.

Relatively, AI is more advance in military and space technologies. The audio, visual, odor, chemical, nuclear and other sensors collects thousand terabytes of information, fusion artificial intelligence analyse and make sense of it in real time to alert army. These tasks perform million times more effectively with old equipment.


If AI progresses in the desired way then the benefits are immense. It can solve complex problems like– prevent disease, understand pattern for earliest diagnosis, path-breaking insight to ailments, increase longevity, double quality outcome at half cost.

Intertwined Big Data and AI: To create the functionality and achieve desired outcome, AI uses big data. At the same time, big data is meaningful because of artificial intelligence. Actually, data makes software translate capability to smartness (perceived intelligence) through huge computational power. Any company, government or entities who generate big data will emerge as dominant AI player.

Steven Hawking warned on AI, “the last invention we’ll ever make and last challenge we’ll ever face”. Elon Musk termed AI threat as “the biggest existential threat” and “summoning the demon”.

Existential threat: Earlier technologies have created new jobs but AI is a net forfeit occupation story. May be only creative, entertainment and emotion oriented jobs will be left for human beings; AI will do the rest. That may force governments to ensure universal basic income.

The Lethal Autonomous Weaponized (LAW) aircrafts programmed to kill without any human intervention in any part of the globe is in operation for military. It raises three ethical questions – Is it right for machines to kill right target on own? What if the machine hits the wrong target? Is delegating the decision to machine to kill human correct?

AI has extremely intrusive power of surveillance and data mining capacity and raises serious concerns of privacy. One week of Super-intelligent machine work will be roughly equivalent to twenty thousand years of human level work; result in disproportionate power and wealth among few. It will be demeaning for humans and will instill inferiority complex.

Any entity six month ahead of competition on super intelligent machine means advantage of thousand years. The advancement in AI will be through time scale and gradual progress towards it will result in disproportionately increase advantage. Therefore, the time continuum of 50 or 100 years to Super intelligence machine does not matter.

Steven Hawking warned on AI, “the last invention we’ll ever make and last challenge we’ll ever face”. Elon Musk termed AI threat as “the biggest existential threat” and “summoning the demon”. Stuart Russel was pessimist too “Nuclear power gave us access to almost unlimited energy stored in an atom, but unfortunately, the first thing we did was create atom bomb. AI is going the same way”. Unfortunately, the terrifying side of AI by Hollywood is becoming a reality.

Opposition of new technology: Historically, new technologies have attracted apprehension of people. Socrates believed that invention of quill and ink for writing would develop forgetfulness in human and they will not attain true wisdom. Printing press was opposed as it could not replicate the secret art of handwriting and corrupt people’s minds with inferior books. Invention of gas lamp post in street were thought to be against the will of God as night is to sleep and Pope Gregory xvi banned it. Locomotive opposed for many reasons – one was to harm human due to high speed. TV was prophesized to ruin society and damage eyes. Computers were to take our jobs but digital revolution has created many new jobs.

Isaac Asimov has shown us the path to make AI benevolent by the three laws of Robotics in 1940; even before robots were in existence!

  1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
  2. A robot must obey orders given by human beings except where such orders would conflict with the First Law.
  3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

Policy, Regulation and Compliance ensure Democratic and Benevolent AI: Fear of unknown make people oppose but the solution is to discuss in public, take preventive actions to stop the misuse and build the confidence. Human level AI is decades ahead and Super Intelligence Machine is further away so we can create protocols and guidelines to prevent AI to be detrimental to human race. Whether AI can assist or exterminate humans, depends on our action. Few suggestions are below:

  1. Mandatory embedding of common good principles in the design of all AIs. AI will evolve through selection process by human; badly designed will be stamped out.
  2. No to unguided research. Transparency of the AI projects.
  3. AI development should be in harmony with human values.
  4. AI should be subject to compliance and audit. AI to hire employee ought to adhere to HR policies and rules.
  5. Compulsory internal ethical committees in companies and research units.
  6. We should have distributed technical capability of AI for balance of power rather unchallenged concentration among select few. Open AI research could be an answer. Inclusive competency will help to overcome badly designed or rogue AI by the others.
  7. Set up field of inquiry committees to ensure safety screening and monitoring.
  8. Strengthen research collaboration among safety groups, technical fraternity, developer groups and other stakeholders.
  9. Strong collaboration among all stakeholders like futurist, academic, researcher, company, public funded organization, government, multilateral institution, people foundations etc.
  10. To avoid oversight, the monitoring mechanisms can be at unit, national and international level.

Defining and establishing legal framework and technological mechanism to investigate wrongs will help in stimulating AI research. Well-articulated mechanisms to investigate and place responsibilities should settle the cases like accident by self-driving car or wrong diagnosis by computer. Accordingly, the license and insurance policy will come into existence.

AI applications and tools will develop capability to perform tasks but the test is to abide by the predictable human protocol, social convention, sophistications and sub conscious rules in real situation otherwise it remains a promise and tech fantasy. If all stakeholders work together then humane general artificial intelligence will be the best invention for humanity.

DISCLAIMER: The opinions expressed in this article are of my own and do not reflect the views of past or present employers. The article does not intend to solicit business, technical or legal advice. One should not act based on the information without seeking appropriate counsel from an experienced expert of the field. I disclaim all liabilities in respect to actions taken or not based on any content of this article. Email sent to me will not establish any commercial relationship and please refrain from sending any confidential or sensitive information.