Artificial intelligence robot with circle, chemical structure and program icon on black background. Credit – Getty Images Yuichiro Chino
We must live in the golden age of science.
For centuries, the scientific method has been defined by two pillars – theory and experiment. Now, we live in the age of artificial intelligence, which adds a vital third pillar. Without advanced accounting, according to leading scientific bodies, the discoveries of the past decade, such as the discovery of the Higgs boson, the discovery of new drugs such as helicin, which can kill strains of bacteria resistant to all known antibiotics, or the observation of gravitational waves, “it would have been impossible.”
But despite these advances, scientific innovation today is often defined by new use cases of existing technologies or the improvement of previous developments, rather than by creating entirely new areas of discovery.
In everyday life, AI is ubiquitous in our homes, from Alexa buying our groceries with a simple command, to Netflix anticipating what will entertain us through the ingenuity of algorithms. But we need a lot of it in our labs — pushing science forward for the public good, helping us solve the toughest problems of our time, from climate change and poverty to healthcare and sustainable energy.
This can only happen by accelerating the next global scientific revolution – by supporting the broad and deep integration of AI technologies into scientific and engineering research. Because while the innovation of AI has been substantial, its adoption in scientific and engineering research has not been ubiquitous, rapid, or interdisciplinary.
Why, despite the remarkable advances in artificial intelligence, does it not yet continually help us achieve the kind of breakthroughs that will expand the frontiers of our knowledge, and speed up the process of scientific discovery?
There are two main reasons. First, while a lot of money is already being poured into AI projects at universities, that money tends to be allocated to specific disciplines, such as AI for computer science, rather than work that builds bridges between the natural sciences, computer science, and engineering.
At the moment, the use of AI tools in the scientific and engineering research ecosystem is still in the early adoption stage, rather than a default part of the researchers’ toolkit. We cannot expect scientists to embrace AI capabilities without proper training. A researcher hoping to use AI will need to gain not only a deep understanding of a particular problem — such as antibiotic resistance — but also knowledge of the data, and what a representation of that data is, that will be useful for training an AI model for the solution. He. She.
Second, there are simply no incentives for young scientists to attempt truly bold research. Much postdoctoral funding is associated with specific research grants and expected outcomes are within disciplinary limits, so postdoctoral fellows usually do not have complete freedom to take risks with new technologies.
So what can be done to change the status quo? We believe that training in AI in science, equitable access to AI tools, and its ethical and responsible application should govern any meaningful response.
First, we need rigorous, multidisciplinary training of young scientists using artificial intelligence. AI failures can largely be attributed to unrealistic expectations about AI tools, errors in their use, and poor quality of data used in their development. Scientists across disciplines, from all backgrounds, will need to master artificial intelligence to prevent such errors.
Postdoctoral research is a particularly opportune moment in a scientist’s professional life to receive this training. This may seem illogical, as traditional academic pressures dictate the rapid publication of postdoctoral papers. A degree is earned, before moving on to the next job. But this is the perfect time to broaden the horizons of research rather than fall into the orthodoxy of hyperspecialization. Rather than rushing to establish themselves quickly, postdocs should be given time and support to try something new.
Second, we have to ensure equitable access to AI tools. According to a recent report by the National Artificial Intelligence Research Resources, equitable participation in cutting-edge AI research is limited by gaps in access to necessary data and computational power. The exclusion of scientists from historically underrepresented and underserved backgrounds “limits the breadth of ideas embedded in AI innovations and contributes to biases and other systemic inequalities.”
We have an opportunity to anticipate and eliminate prejudices rather than deepening and strengthening them. We hope that through our philanthropic efforts, by expanding access to AI tools to a generation of postdoctoral candidates around the world over the next several years, we will be able to lay the foundation for fair AI.
Third, the responsible application of AI should enhance human intelligence, not replace it or repeat its mistakes. The power of AI in science is just beginning to take off, but we have to remember that breakthroughs like the discovery of halicin could not have been achieved by humans or only AI. There is clear evidence that AI can increase the analytical ability of humans, and perform complex experiments beyond traditional methods. For example, research published in Nature shows how an AI algorithm can help drive a promising path to sustainable energy by containing and controlling high-energy plasmas for fusion energy research. AI can also discover theories at the forefront of mathematical research.
But the real excitement in the application of artificial intelligence to science is in the new areas of research that we cannot yet perceive, which will bring new dimensions to the history of the scientific method. The microscope made it possible to examine a completely new world of microorganisms that the first biologists had not thought of. The telescope showed early astronomers just how vast the universe is outside our solar system. AI can help us discover new phenomena that human scientists have not yet considered.
Changing the way we use science is not primarily about artificial intelligence. It’s about people and the wisdom we bring to technology. We have no doubt that with the right support, early-career scientists are ready to embark on a wave of new discoveries: more effective drugs, renewable alternatives to plastic, sustainable energy production and storage, and deeper insights into our world and our biology.
schmidt They allocate $148 million to fund new Eric and Wendy Schmidt Prize for Artificial Intelligence in Postdoctoral Sciences, a program of Schmidt Horizons. This will support Hundreds of young scientists and engineers at nine universities around the world to drive innovative, interdisciplinary use of artificial intelligence in science, technology, engineering and mathematics.