Mankind’s Last Invention — Superintelligence, Singularity and Intelligence Explosion

With the latest advancements in Artificial Intelligence, a key term which is resonated throughout media these days is “AGI” or Artificial General Intelligence. Tech podcasts and blogs are throwing around this term in such a fashion, implying the impending revolution mankind is about to face next. Today we will take a deep dive into what AGI is, how it will happen and what its impact on human society will be like.

For readers who are unfamiliar with AI technology, I will first explain what AGI is. The domain of AI is categorised into three types based on capability; Narrow AI, General AI and Super AI. We currently experience Narrow AI in our devices and apps; an artificial intelligence software catered to a specific task such as image classification or transcribing audio. In contrast, General AI can understand and learn any intellectual task that a human is capable of. This means a single software that can learn to do anything given enough resources; it can learn to trade stocks or draw visual masterpieces. AGI has not been invented yet, although research is carried out extensively in pursuit of this holy grail of machine intelligence. Experts point out that if we were to someday craft the first AGI, we are bound to experience several cataclysmic events. One such alarming effect is Technological Singularity.

What is a Singularity?

My own definition of a Technological Singularity is,

A significant point in time where technological advancements occur at a much faster rate causing severe contrasting differences in society and civilisation.

Bob is an entity from the 1700s. He is familiar with the fairly new traditional railways, a couple of motorcars on the road and the occasional telephone call. Suppose we bring Bob to 2023 with the help of a time machine. Much to his surprise his environment has taken a drastic change. Monolithic skyscrapers reaching the sky, hundreds of dashing motor vehicles of all shapes and sizes flowing through massive highways spread towards eternity, and slim handheld devices to which every person is glued. To Bob, this society is unrecognisable. This transcendence from the industrial age to the information age can be described as a singularity.

In this definition, singularities have already occurred throughout history. The invention of the wheel was a singularity event in the stone age. Egyptians invented the pulley, revolutionising mechanical advancement. The steam engine and the telephone were pioneers of the industrial revolution. Then we have the Silicon transistor, satellite communication and the smartphone which are pillars of the Information Age, today.

But what new technology will carry the future of civilisation? Some professionals declare that Nanotechnology will be the key to technological nirvana. Nanotechnology is expected to aid in feats such as curing cancer and evolving computation power. Other experts confirm that technologies like Quantum Computing will be our saviour. Solar-powered electricity generation, biotechnology, and genetic engineering are other sciences under development which may elevate mankind to a tech utopia. But in this article, we will address the elephant in the room, AGI. Artificial Intelligence in its current state of development has already affected many industries; the trending product advertisement in your Instagram feed or the next Netflix show you hope to binge-watch are insights drawn by rigorously trained AI. But what awaits us in the future makes recommender systems and AI-powered medical diagnosis seem like stone age technology.

Superintelligence

While AGI performs as same as a human mind, super AI or superintelligence outperforms human thinking and computation. There are several forms of superintelligence put forward by experts but I will explain only two forms in this article.

There is a possibility, with the advancements in bio-medical research, to integrate computer-powered systems into human anatomy. Companies like Neuralink are already working on this domain as they believe we can surpass our natural biological limits through such technology. While that kind of a future can be conceptualised as a cyberpunk society with everyone being a cyborg, human-machine integration may be leveraged to significantly enhance our brain power allowing human brains to perform faster and more powerful computations. This can be classified as human superintelligence.

Although human superintelligence is possible, the more plausible and economical form of superintelligence is speed superintelligence. Here, a machine is first taught to think like a human (AGI). In contrast to the human mind, such intelligence runs on powerful processors and has much more computational, simulation and storage power. Thus, such intelligence will easily outperform us humans in speed, making it the most intelligent entity on Earth. But what will happen, if we were to let it improve itself?

Intelligence Explosion

Imagine the first AGI has been developed, it runs on IBM’s most powerful Quantum Computers, making it superintelligent. The developers flip a switch, which enables the AGI to edit its own code and make improvements to both its hardware and software. The research team then secures the connection of this system to the internet and waits. Upon several hours of iterations of the program, the AI has tripled its calculation speed, minimised its memory utilisation and created new patterns in reasoning. It now improves its intelligence exponentially, giving birth to a key event called an Intelligence Explosion. Below is a definition for intelligence explosion proposed by British mathematician I. J. Good in 1965.

“Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. Since the design of machines is one of these intellectual activities, an ultraintelligent machine could design even better machines; there would then unquestionably be an ‘intelligence explosion,’ and the intelligence of man would be left far behind. Thus the first ultraintelligent machine is the last invention that man need ever make, provided that the machine is docile enough to tell us how to keep it under control”

No surprise, with such a compounding intelligence (Imagine the IQ of this machine being doubled every 2 hours) human intelligence would be left behind making us seem incredibly idiotic. Such intelligence can work for a multitude of society's problems to disappear. The cure for cancer, solution for poverty, the establishment of world peace, state-of-the-art algorithms and all sorts of solutions will emerge rapidly. Nobel prize-winning discoveries will be put forward every few seconds. The advancement will be so drastic that we humans would not be able to keep up with it.

But When?

The exact time period when the next singularity will occur cannot be predicted. But there are numerous professionals and experts putting forward their expectations for AI to take over the intelligence race. Elon Musk predicts superintelligent AI to cause a singularity by 2029, that’s soon. While some others debate that it will happen in the 2030s or 2040s, some even argue that it may never happen at all. By the time I am writing this article, my personal belief stands that humans will one day invent AGI and my estimate is within the next 50 years.

This blog post sounds much more like science fiction rather than an article, yet with the number of ground-breaking discoveries we had in AI within the last year, the idea of AGI emerging cannot be disregarded. And if crafted one day, its impacts on society cannot be taken lightly. If an intelligence explosion were to occur we would have to prepare for its consequences, as a new species of intelligence is likely to emerge, setting us one level below in the dominance hierarchy. The possibility of AI turning against humanity, as fantasised in movies like Transcendence (2014), is debatable. AI experts refer to this problem as Alignment and is a topic for another discussion. But the technological change we have to undergo is imminent, so brace yourselves.

“Intelligence is power”