Holiday Flash Sale!
Artificial intelligence can improve health, protect biodiversity and even write wine reviews
Artificial intelligence does not have to threaten humans; it can also work with us to solve big problems. Are you still feeling a little wary of algorithms? We rounded up a slew of stories from the past year that demonstrate the many ways in which this technology can have a positive impact.
This year AI revealed its prowess as a powerful tool to help prevent climate change from wreaking irreversible damage to the planet, something that requires more than one solution. Researchers have been using AI to visualize the future effects of floods and wildfires, improve climate decision-making, monitor forests and share data. Other climate projects powered by artificial intelligence have included building digital twins of the planet to test out the impact of different warming-mitigation policies and mapping thinning sea ice.
Without access to a variety of foods, people around the globe are at risk of malnutrition and other health problems. A new model pinpoints the locations where populations are more likely to suffer from nutrient deficiency. In many U.S. cities, meanwhile, access to clean water is hampered by the presence of toxic lead pipes. A new algorithm helps local governments find and remove these underground menaces by predicting which homes are most likely to have them.
Buried land mines kill thousands of people each year. But researchers are working on a system that uses drone footage and machine learning to pick out the unexploded munitions from a distance so they can be safely disarmed. Of course, drones themselves can also be used for nefarious purposes—but yet another AI algorithm, this one inspired by the eye of a fly, can detect these threats.
An AI program defeated human video game champions at the ultrarealistic racing game Gran Turismo. This will help the game’s developers give competitors a worthy automated in-game adversary. And beyond that, self-driving car researchers could use the program’s success to inspire their own work in the real world.
This year researchers completed a massive field test of an AI program for detecting sepsis, one of the leading causes of death in hospitals. The results suggest that the program reduced health complications and deaths from sepsis while earning positive feedback from the medical professionals who used it. And human diseases aren’t the only ones AI can combat—another algorithm has been used to diagnose bacterial infection in olive trees.
Every function in the human body relies on proteins, which consist of long chains of amino acids folded into complex structures that carry out the work coded by our genes. But predicting what shape that chain will take is tricky and can require years to resolve—if you’re a human. This past summer Google’s company DeepMind announced its AI program AlphaFold had predicted the molecular structures of just about every known protein. That’s about 200 million estimated protein shapes. This massive achievement solves one of the knottiest problems in biology, and it won the AI’s creators a $3-million Breakthrough Prize.
Want to preview the experience of sipping a dry white wine or a fruity sour beer? A language-generation AI can help you decide what to drink by combining existing reviews into condensed, summarized descriptions. And its creators say the program could be expanded to distill reviews of a whole lot of other products.
As cold weather sets in, COVID cases are on the rise again. To figure out what the coronavirus that causes the disease will do next, researchers are using an AI tool to analyze viral mutations, predicting when a new variant such as Omicron will emerge and become dominant. Such prediction tools can do more than help manage the current pandemic. Other algorithms are programmed to examine viruses currently spreading through the animal kingdom to identify which ones might jump to humans, potentially helping researchers avert the next pandemic.
Novel designer drugs can produce effects similar to known recreational drugs, but they have small molecular differences. Because the designer versions are chemically distinct from the original drugs, they can sidestep some of the governmental regulations on these substances. They can also have unexpected debilitating impacts on health. So many molecules have intoxicating potential that banning these substances has required regulators to play a game of whack-a-mole—but now AI has stepped in. Researchers used an algorithm to produce a database of molecules that could be developed as potential designer drugs. It could allow governments to preemptively outlaw these dangerous substances.
There is a tendency to think of technology and nature as forces in conflict. But it doesn’t have to be that way. Artificial intelligence algorithms can analyze data from at-risk ecosystems to measure the biodiversity of that environment—and to support conservation projects run by Indigenous groups. Another AI-powered technology, facial recognition, is helping researchers monitor mountain lions without disturbing their solitary routines.
Sophie Bushwick is an associate editor covering technology at Scientific American. Follow Sophie Bushwick on Twitter
Seth Wynes | Opinion
Ida Emilie Steinmark
Corbin Hiar, Carlos Anchondo and E&E News
Discover world-changing science. Explore our digital archive back to 1845, including articles by more than 150 Nobel Prize winners.
Scientific american arabic
© 2022 Scientific American, a Division of Springer Nature America, Inc.
All Rights Reserved.
Support science journalism.
Thanks for reading Scientific American. Knowledge awaits.
Already a subscriber? Sign in.
Thanks for reading Scientific American. Create your free account or Sign in to continue.
See Subscription Options
Continue reading with a Scientific American subscription.
You may cancel at any time.
Holiday Flash Sale!