A new study indicates that artificial intelligence not only optimizes the work of scientists, but also changes the face of scientific research. The results: a jump in discoveries, patents and new materials - but also challenges in the researchers' sense of satisfaction and happiness
The futurist Ray Kurzweil has been promising us for years that we will still have eternal life, perfect health and so on. At the same time, he admits that we are not even close to the scientific knowledge necessary to achieve all of these. So why is he so sure that we will still reach that period, and even in the coming decades? Simple: he trusts the artificial intelligence to boost our scientific research capabilities.
But can she really do it?
A new study released in early November Shows that artificial intelligence is definitely able to jump-start scientific research. Along the way, he also teaches us some important lessons about the future of work - at least in the coming years.
The study was conducted on 1,018 scientists who work in the research and development of new materials and products in a large American company. In the last two years, the scientists received access to artificial intelligence that offered them 'recipes' for new materials, with interesting chemical and physical properties. The scientists had to try the recipes, develop the materials and test their properties. Once such a useful new material is developed, it is incorporated into the company's existing products.
This is an important point: it means that materials developed with the help of artificial intelligence are already in the products we use. If you thought artificial intelligence was only used to write a love song for your cat, well, no. Companies are already using it today to change the world. In the coming years, almost every product will be re-examined to see if it can be improved using artificial intelligence.
But for now we will return to the present.
In a study conducted on the researchers' research in the research laboratories, the researchers saw that the researchers who were helped by the investigating artificial intelligence, discovered 44 percent more new materials than their peers. Those new materials also had impressive physical properties, so artificial intelligence not only increased labor productivity, but also the quality of the products. A similar increase was also recorded in the number of patent registrations, and a few months later it was possible to see a 17 percent increase in the number of products (which were still being tested in the laboratory) that contained the new materials.
These numbers are impressive, but it turns out that this is just the beginning. A significant number of scientists have barely managed to benefit from AI, while others have almost doubled their output thanks to it. What differentiates the two types of scientists?
The difference lies in one critical attribute: expertise. But not expertise in the use of artificial intelligence, but rather in the field of research itself. And of course, in critical thinking.
The real experts knew very well the materials that the AI suggested to focus on. They knew when she was 'talking nonsense', and which ideas were more successful. Their friends, the ones with the low expertise, jumped on every idea of the artificial intelligence, and fell time after time. The experts, on the other hand, critically and carefully examined the directions that the artificial intelligence suggested, identified the ideas with the greatest potential for success, examined them in depth - and improved their productivity by 81 percent.
When the company laid off three percent of its employees in the last month of the study, do you want to guess who was laid off first? True: 83 percent of the scientists who were fired were those who failed to benefit from artificial intelligence. The positive side is that the lab recruited a larger number of new researchers. And if I had to bet, I'd put good money that those new scientists were tested on their ability to get the most out of artificial intelligence. That is, they had to be experts, critical - and yes, also know how to work with artificial intelligence.
We see here how the way of working in laboratories changes to adapt to the new world and the new technologies. Most people experience such changes as a shock. We don't like to change too sharply, or realize that the previous way we used to work is no longer relevant.
This is what also happened in the study here. The researchers did significantly improve their productivity, but they also enjoyed it less. According to satisfaction surveys, they experienced a 44 percent drop in job satisfaction. The decrease in satisfaction can be found both among the "losers" - those who failed in working with artificial intelligence - and among the "winners". All indicated that they felt that their skills were not well utilized, and that they required less creativity in their work. They were happy with the increase in output, but it just wasn't enough. Overall, almost all of the researchers experienced a decrease in their overall sense of well-being.
And they realized they had to rethink their work.
Working with the new tools changed the way scientists thought about artificial intelligence. Their belief in AI's ability to improve lab productivity almost doubled. At the same time, they realized that it was going to change the skills required to succeed in their field, and the number of researchers who stated that they intend to acquire new skills and ways of thinking has almost doubled since the study began. When the research just started, very few of them thought that they would have to acquire new skills. When it was finished, two years later, it was already a fait accompli.
Lessons for the future
It is difficult to find broad and impressive studies like this, which rely on evidence from the field and close monitoring of the workers and their products. The conclusions seem clear: the correct integration of artificial intelligence in the research work, can increase the output almost twice. But for that, employees need to know how to use AI and how to critically examine its ideas. And this they can only do if they have a high level of expertise in their field.
This means that, at least for the time being, the experts among us can be confident in their work, but they must understand that if they do not work with artificial intelligence - they will be left far behind.
At the same time, employers should begin to consider the implications of the integration of artificial intelligence on the employees' sense of competence, enjoyment and self-satisfaction. It is now clear to everyone that artificial intelligence can take over an increasing share of tasks that were previously considered only human, and humans are starting to get stressed. What is the meaning of our existence in a world that does not need us to the same extent? How can we justify to ourselves the great effort we invested in acquiring skills that are no longer needed? These are not questions about a gloomy and distant future. These are questions, doubts and difficulties that employees raise here and now. Already today. We feel threatened by artificial intelligence, and the psychology of employees is affected accordingly.
I don't want to ignore the negative impact AI will have on some workers. Yes, many will have difficulty finding their new way in the first years. It's sad, but that's what happens in every scientific, technological or industrial revolution. I say these things from a very personal place: I myself see and feel how some of my skills are no longer relevant to the new world. But sadness is not a solution. I am sad - and I choose to move on, acquire new skills and find new interests. This is the way of the world, and this is the way of progress. We will all have to adapt ourselves to the future, instead of expecting it to adapt itself to us.
Still, I want to end this post with some optimism. Actually, with a lot of optimism, because a lot of good will still emerge from the integration of artificial intelligence in scientific research. In just two years, artificial intelligence almost doubled the productivity of the scientists who knew how to work with it. These are amazing results, and this is just the beginning, because we will still develop more sophisticated artificial intelligences and learn to use them more correctly in the research laboratories. The rate of improvement in scientific research is only expected to increase. In the coming decade, expect a surge in the number of new scientific discoveries, new medical insights, drugs and new developments and patents.
Suddenly, Ray Kurzweil's prediction of eternal health and living as long as we want doesn't seem so delusional anymore. True, in order to realize it, we need science to improve by leaps and bounds, at a pace unprecedented in human history.
And here it happens.
To the eternal life!
More of the topic in Hayadan:
More of the topic in Hayadan:
