AI Is Officially Born. What Will We Do With It?

Author(s): Valeria Fonseca Diaz

Originally published on Towards AI.

I heard a pretty good joke from John Oliver in his episode about AI where he told the story of the Atari game Breakout. The computer learned and created strategies to play the game, winning flawlessly. Then he goes on saying, “Of course it got good, it did literally nothing else!”.

Deep learning came a few decades ago and revolutionized without mercy the human mind and its intelligence. We could have predicted it from the invention of the calculator. If such a device was already able to do basic calculations much faster than the human brain, it was probably obvious that human intelligence was going to be surpassed by computers sooner than later. Even if we knew it back then, AI’s progress is one big unprecedented revolution in the first quarter of the 21st century.

Back in the 19th and 20th centuries, one of the biggest disasters in the history of humanity was nuclear weapons. Einstein’s discoveries were so revolutionizing for how we understood the structure of the universe that he completely changed the direction of the history of science and opened the doors for so many new theories: String theory and black holes, for example. He was likely amused by being the brains behind this scientific discovery. Never would he have thought that others could have the intelligence to use his theories to create something that could literally destroy the entire world. The discovery that all energy can be converted into matter and all matter into energy was the simplest theoretical fact that gave rise to such deadly weapons.

Besides the creation of the atomic bomb, silent, Baby boomers, and even Generation X, faced many more dramatic events and revolutions in human history: World wars, epidemics, famines, and many more. While serious economical crises arose from those events threatening the survival of humans, all of those events had in common the fact that they were not normalized in the big picture. There were governments committed to defeating those events to bring back safety to their people. For whatever of those events, there was always faith and hope that things would not go too bad in the future and that our governments would be able to restore human peace. This is certainly not the case for the entire human population, but the idea that those catastrophes were not beneficial and that we needed to defeat them was surely living in our species’ minds.

Now, as of 2023, the young generations (those born after 1980) are starting to be part of another kind of revolution in human history. Those of us, though, will undergo this revolution in a very different way from the events that previous generations had to experience. The evolution of AI is not a catastrophe like nuclear weapons are. AI is not an event that will put us in an economic crisis from which we can only hope to survive until things come back under control. As Stuart Russell puts it in his book “Human Compatible”, AI can be the surpassing of human capacity, and “gaining access to considerably greater intelligence would be the biggest event in human history” [1]. This is how we, the young generations, will officially be put through the highest pressure to remain valid for the several decades ahead of us. This is how we, the current young ones, will experience the greatest event in human history, something that, before scaring us, amazes us, while at the same time silently threatens us.

The recommendations that AI gives us for human behavior and decision-making are all based on human data. For whatever we do, the machine will be faster at deciding based on what we have done before. This fact has a major drawback. If the patterns that the machines are trained with are those leading to unfavorable human behavior, we will continue generating the same output with AI with some differences: It will lead us to the result faster and it will uncover the patterns that might have been much less obvious in the past. So, what is spectacular about AI recommending us something is its ability and speed to process the conditions to cause a certain output. Even the creative models, those that come up with their own output, are built with our data.

This shocking progress in automation and artificial intelligence leads to questions about the real necessity of these inventions. Do we really need machines to think faster than humans do? Do we really need robots to do the tasks that humans do? Do we really need this new artificial species to replace human abilities? The answer is, it depends on the task, the type of thinking, and the ability we are talking about. We know that tasks that humans might boring and repetitive, while being time-consuming are something we are happy to give up. The question is, are we equally happy to give up on thinking, learning, discovering, and creating?

We can highlight this contrast with an example: discovering the causes of rare diseases vs. making a comedy or producing music. Both of these tasks are very easily handled by AI with the current progress. AI can uncover so many patterns leading to a concept that it essentially makes no difference whether we use it for one purpose or the other. This type of intelligence knows what our favorite jokes are and what types of sounds humans find more attractive. It can also easily know what causes rare diseases. This one might still not be as clear just because of lack of data, not because of lack of intelligence. But assuming both tasks are equally available, there is a significant impact on human well-being by both. While knowing the causes of a certain disease saves us, we may not be as happy replacing our creative brain. That’s the game we’ve put up now.

As of today, the very idea of AI evolution has come as such a powerful human invention (or discovery), that the consequences of its existence and domination over human activities and minds remain in question. While a nuclear weapon clearly shows us how it can destroy our species, AI shows us first secrets that we could not see before and creates ideas that we would have not imagined before. The amusement of creating or discovering this superhuman intelligence is too high for us to think and see the dangers and future of the human species.

This invisibility of the consequences of AI is inevitably accelerating governments' actions. In a way, we are just waking up to see the progress that has been happening since the 1990s when we still thought that we were far from anything like today’s AI. But AI will continue to evolve at a much faster pace than the pace of our governments’ decisions. Italy banned ChatGPT in March 2023 by restricting OpenAI’s access to their data. Some thought that it was too much of a strict measure and OpenAI was certainly disappointed at this decision. But then, in a matter of a couple of months, we see the impact of ChatGPT after its launch. Companies are now hiring people according to their ChatGPT skills and firing those who are not up-to-date with the tool.

For years, there have been great efforts into defeating inequality around the world. While the US and China are taking over the growth of AI with spiked revenue after the launch of many platforms, so-called developing countries still face very difficult economical challenges, child malnutrition, water scarcity, and the list continues. AI will inevitably bring a much bigger gap in societies and their access to basic needs. In its series of books best-sellers, Yuval Noah Harari discussed in depth the advance in technology and how this could represent the evolution of the human being and the high risk for a larger gap between the rich and the poor [2]. It has only been a few years since his analysis was published and here we are today, seeing the realization of this perspective.

This debate simply brings us to yet another human event after facing the Covid19 pandemic in 2020 and the start of the Russian war against Ukraine in 2022: AI is officially born. And with it, we need to understand what this means. Ideally, AI has come to make the world a better place certainly not only for humans but for everything that depends on it. Ideally, we can empower this type of intelligence to help humans see and uncover the secrets of the world that are invisible to us and beyond our capacity. Hopefully, we dominate AI as a tool instead of making it an evolutionary event that takes over the world. Sure, we can use AI to create new shows, jokes, and art. But we can better use it to eradicate diseases, understand climate change, defeat famine, and so on.

AI could be a way of working, rather than being an objective in itself. We can become smarter at a much faster pace than we would without AI. In the future, we do not need to fear losing our jobs and staying behind the evolution of AI if we see it as a tool that empowers us instead of dominating us. This path will lead us to shape our way of thinking and working not blindly using AI, but using sustainable AI. Current governments’ decisions will certainly democratize this concept.

While we continue talking about the evolution of AI and its impact on humans, we need to start preparing for an emerging AI culture. Not only do new trends emerge in a fraction of a second now, but normalization processes will run much faster for new concepts and habits. AI will not give a chance to human resistance towards change. We will need to become much more flexible in adapting. Perhaps this will be the greatest human trait we can develop in the AI era that has begun.

How can we learn to adapt faster? That’s our biggest inquiry.

[1] S. Russell. Human Compatible: Artificial Intelligence and the Problem of Control. Penguin Publishing Group, 2019

[2] Harari, Y.N. Homo Deus: A brief history of Tomorrow. New York, NY: Harper Perennial, 2018

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI