Branded content people actually care about
Branded content people actually care about
What's Next In TechWhat's Next In TechFrom robotics to health care to TVs and more, Best Buy and Studio@Gizmodo are exploring the future of technology.

When most people hear “artificial intelligence,” or “AI,” they tend to imagine the kind of AI depicted in sci-fi entertainment: cyborgs, sentient robots, and other human-made machines that end up outsmarting their creators.

Illustration for article titled Why AI Might Not Be The Boogeyman We’re Dreading

But what most people mean when they talk about AI is really machine learning — a branch of AI that comes down to math. Machine learning involves analyzing data and applying a statistical model to make predictions for the future: songs you’ll want to listen to, what products you’re likely to buy, or the probability you’ll be late on your next credit card payment. The more data they’re fed, the more these models “learn.”

This kind of AI gets a bad rap in the press, and it’s easy to understand why. AI now touches nearly every part of daily life — from the prices shown on rideshare apps to content selected for social media feeds — and are shaping how humans view and interact with the world. Automation and facial recognition are being implemented in ways that can have worrisome impacts on society.

There is nothing inherently bad about AI, though — and plenty of innovators are working to use it for good. It has the potential to tackle problems, move society forward, and expand the well of human knowledge. This notion strikes at the heart of What’s Next In Tech — our partnership with Best Buy where we’re exploring how innovations are improving our lives. Here are some of the most inspiring and beneficial ways that AI is being used today.

Game-playing has preoccupied computer scientists since the profession’s inception in the mid-20th century. Alan Turing — considered by many to be the father of computer science — hatched the idea of a computer program that could play chess as early as 1950. In 1997, IBM’s chess-playing computer, Deep Blue, beat chess champion Garry Kasparaov. Twenty years after that, in 2017, an AI program called AlphaGo beat the world’s best Go player, Ke Jie, in three games.

AlphaGo inspired a new generation of computer scientists to pursue AI and to tackle games. Stephen McAleer, a PhD student at the University of California at Irvine, became interested in machine learning after reading about the program’s victory. This year, McAleer and several fellow researchers, including Forest Angostelli, developed a deep-learning problem that solved the Rubik’s cube with minimal help from humans — a feat that nobody had accomplished before. “The same algorithm is able to solve other hard combination puzzles, such as the 35 and 48-puzzle, Lights Out, Sokoban, and Towers of Hanoi,” McAleer explains. “The hope is that it can be used to solve any large planning problem where you know what the goal is, but there are too many states to use classical planning algorithms.”

Possibilities beyond puzzles and games include the prediction of tertiary protein structure — an application that’s important for medicine. “It is the responsibility of [AI] researchers and practitioners to apply their skills to tasks that will benefit humankind,” Angostelli said.

Around the world, various teams of researchers and medical experts are working to train algorithms to detect cancer. The premise is similar to facial recognition: Doctors annotate and label X-rays, CTs, and other medical images to make features recognizable to computers. Even though the fight against cancer has yet to be won, it’s promising for health care overall. “You can use this technology to improve the consistency of diagnoses and save doctors time,” says Lingge Li, a statistics PhD student who has spent his summer researching at a large biotech firm. “It can also make health care more accessible because you can do this remotely.”

Illustration for article titled Why AI Might Not Be The Boogeyman We’re Dreading

Li’s research extends to using machine learning for drug discovery, another health care field that has significant potential to benefit from AI. Clinical trials are costly, and most don’t succeed, Li explains. With AI, experiments can be performed in vitro rather than on animals and in humans. “You can see how different chemicals interact and use machine learning to select the most promising candidates,” Li says. The process could catalyze the discovery and development of life-saving drugs.

On the patient side, AI is precipitating a positive impact too. A start-up called Univfy is using machine learning to help women who are trying to conceive. The company, which was founded by two Stanford professors, uses AI to predict accurate outcomes of in vitro fertilization with the goal of eliminating the emotionally draining uncertainty that often accompanies IVF cycles. Users’ personal data aren’t beamed to a centralized location, either — the company works with clinics individually, and the model uses only the local patient data kept by each clinic.

Scientists working on conservation are turning to AI for everything from monitoring wildlife, to tracking ocean pollution, to measuring the effects of natural disasters on landscapes. “Together with advances in satellite imaging technology, there is great potential for AI to help us monitor and protect our natural resources,” says Peter Sadowski, an assistant professor in computer science at the University of Hawaiʻi at Mānoa. Sadowski’s team is working with oceanographers to automatically detect important features in high-resolution satellite images of the ocean — from debris, to sea ice, to different kinds of waves — all over the world, every six days.

At Cornell, meanwhile, researchers are working to use AI to protect African elephants from rampant poaching. Elephants in the region are notoriously difficult to track due to thick forest — but thanks to a machine learning tool developed by Conservation Metrics, scientists are now able to identify the low frequency of elephant calls amidst the cacophony of rainforest sounds. Using that data, the organization hopes to better map elephant habitats and learn whether elephants are traveling in areas that are especially vulnerable to poaching.

In terms of human impact, the potential of machine learning is just as profound. Much like its role in drug discovery, AI can be used to help evaluate and discover more efficient resources, from construction materials to fuels. It can be used to optimize processes that take a toll on the environment, such as shipping. And it can be used to better things like predict weather and energy consumption.

Sweeping societal advancements aside, it’s also important for people to have fun, and to improve and enjoy their daily lives. Artists are exploring using AI to create new experimental works, like music from an algorithm trained on the composer’s singing voice, and paintings created from AI trained on a body of artworks. In 2018, the first AI-created painting sold by a major auction house — a male portrait with an unfinished appearance, based on a dataset of pre-20th century portraiture — was snapped up by an anonymous bidder for $432,500. That doesn’t mean that AI is rendering artists obsolete: Human artists are still behind the code that’s written to create these works.

Studies have also shown that interactive environments help students learn languages — and students at New York’s Rensselaer Polytechnic Institute are learning Chinese through an AI-powered virtual environment that teleports them straight to China, where they can practice their Chinese with street vendors.

In the entertainment realm, AR and VR — which depend on computer vision, a field of AI that trains computers to process visual information — are becoming increasingly accessible and developed for consumers, with hundreds of VR games, including fitness games that can be connected to exercise bikes, available on Steam and Playstation. Other indie video game developers are exploring the use of AI to create more intelligent games with more realistic graphics, and perhaps eventually, games that are tailored to each individual player’s gameplay.

While we’re already seeing amazing advances in AI implementation in console and VR gaming, it’s exciting to think about where the technology is headed, and how it might make its way into products available at Best Buy in the future. For more incredible breakthroughs and advancements in tech, check out our What’s Next In Tech special section.

Angela Wang is a freelance writer living in Queens.

This post is a sponsored collaboration between Best Buy and Studio@Gizmodo

Share This Story

Get our newsletter