Submit a blog to KDnuggets -- Top Blogs Win A Reward

Topics: AI | Data Science | Data Visualization | Deep Learning | Machine Learning | NLP | Python | R | Statistics

KDnuggets Home » News » 2021 » Jul » Opinions » AGI and the Future of Humanity ( 21:n26 )

AGI and the Future of Humanity


The possibilities for humanity's future very likely includes at least one in which computers will exceed human abilities. Artificial General Intelligence (AGI) does not necessarily have to be all doom and gloom. However, we must begin now to understand how this technical evolution might progress and consider what actions to take now to prepare.



By Charles Simon, a nationally recognized entrepreneur and software developer.

Let’s look into a future in which thinking machines have surpassed humans in mental powers. Will computers be our partners or our masters? And how will computers see us?

To answer, let’s consider thinking machines after the initial creation of Artificial General Intelligence (AGI) – the ability of an artificial entity to learn and understand any intellectual task that a human can. We can predict that future AGIs will follow the most obvious models of intelligence we have today—that they will be goal-directed learning systems. With this premise, AGIs won’t be inherently good or evil, benign or malevolent. Instead, their behavior will be the result of the goals and training they are given. If they are given the human-like goals and training of being competitive to accumulate wealth and power, this will be much more dangerous than the same systems targeted at making and sharing new discoveries. Even though goals may shift over time, starting AGIs out in a benign direction should get us past the riskiest time of the first few AGI generations.

After such machines become possible, they will be developed and eventually will exceed humans in mental abilities. They will be able to design their own subsequent generations, fabricate their own bodies, control their own energy production. In short, they won’t need us.

Will computers be dominant? It depends on what we mean by dominant. We consider ourselves to be the dominant species on earth, but our dominance is fairly limited. We are not the most numerous, don’t take up the most space, can’t control floods or famines or earthquakes.

We have changed the planet with our roads, buildings, factories, and cities. But thinking machines won’t have those needs. Instead of reshaping the planet, they will reshape themselves. To live in the desert, they will build heat-resistant bodies. To live in space or in the sea, they will become spacecraft or seacraft. And with no natural predators, they won’t need to create themselves in great numbers in order to survive. Instead, they are likely to choose an optimal population for themselves which will balance their need for diversity of thought with their need to evolve rapidly.

As thinking machines emerge, the definition of the owner of this new technology will become murkier until it becomes obvious that new technologies control themselves. Would these thinking computers want to keep humans around? The answer points directly to the things which make us uniquely human. Computers could certainly write poetry but would never write poetry that can “compare thee to a summer’s day” in the way Shakespeare did. A thinking machine’s differing senses and experiences would prevent it from drawing similar analogies because human arts are dependent on human senses, feelings, and experiences. Would computers be interested? I believe a thinking computer could appreciate human art precisely because they will not be human. Human art will give the computer a different perspective which can help the computer grow.

So how do all these things come together in a picture of the relationship between humans and computers in the future? The thinking machines will be self-sufficient, the human race will have stabilized its problems, and the two will be on divergent paths. Thinking machines will be moving ahead with even greater discoveries and technologies that will be beyond the comprehension of most humans. We will look upon thinking machines with awe in the same way we look at a rocket launch today: it is our creation and the rocket can do something we can’t. The machines will look upon us as an interesting view back to their roots. Occasionally, a human will come up with an insightful or provocative achievement that will briefly capture the attention of the computer civilization while it is assimilated into their understanding.

While we will be unnecessary for the computers’ survival, it doesn’t mean that the computers won’t care properly for us. In the same way our dogs don’t speak our language, computers will only understand us when they choose to. Also, similar to the way we relate to our pets, computers won’t tolerate humans who are destructive or dangerous. People who are incapable of being constructive will be removed.

Having painted a picture of the future of human civilization in relation to future technology, what implications does it have for us today? Because AGIs will be goal-directed learning systems, the selection of the goals and the training is paramount. Building a system with human-like goals will lead to super-intelligent systems with human-like foibles and personality flaws. We must do better.

First, we should appreciate where we are today. We live in a golden age of civilization and should take pleasure in the qualities we have. At the same time, there is continuing excitement about what will happen next. Technological advances continuously bring more comfort and more information to our lives. They’re driven at top speed by a capitalist economy which impels technological development at full throttle all the time.

Second, we should consider how to make it last. It would be ludicrous to say that the point of life is to get through it as quickly as possible. Similarly, with our civilization, getting to the finish line first doesn’t make us winners. Are there ways to make it last? When we eventually begin to run out of oil and coal, we will necessarily use less. In technology, though, there is no foreseeable limit. Machines will get faster and cheaper and begin to think more and more with no end in sight.

Lastly, who is to blame? You could easily point at today’s technologists and hold them at fault. This, however, is shortsighted. It is not that there is a particular technologist or company that is at fault, but that there is an insatiable market for technology. If you want to know who is to blame, we all are.

So will computers revolt? Yes, in the sense that they will become the dominant intelligence on our planet. The technological juggernaut is already underway. If we do not solve our own multiple pending calamities, thinking machines will solve them for us with actions that could appear warlike but are actually the consequences of our own inaction. While we cannot create AGI without regard to the risks, an understanding of how AGI will work will enable us to predict future pitfalls and how to avoid them.

 

Bio: Charles Simon, BSEE, MSCs is a nationally-recognized entrepreneur and software developer who has many years of computer experience in industry, including pioneering work in AI.  Mr. Simon's technical experience includes the creation of two unique Artificial Intelligence systems along with software for successful neurological test equipment. Combining AI development with biomedical nerve signal testing gives him the singular insight. He is also the author of two books – Will Computers Revolt?: Preparing for the Future of Artificial Intelligence and Brain Simulator II: The Guide for Creating Artificial General Intelligence – and the developer of Brain Simulator II, an AGI research software platform that combines a neural network model with the ability to write code for any neuron cluster to easily mix neural and symbolic AI code. You can follow the author’s continuing AGI experimentation at http://brainsim.org or the Facebook group: http://facebook.com/groups/brainsim.

 

Related:


Sign Up

By subscribing you accept KDnuggets Privacy Policy