4 ways AI is transforming the physical world 

Leaders from Waymo, Exelon, and Hopkins explore how AI is changing transportation safety, energy affordability, and even sports analytics.

Virtual assistants, search engines, rideshare ETAs — all of these common tools leverage artificial intelligence in the digital space, but AI’s applications also extend to the physical world. At the inaugural World Changing Ideas Summit, hosted by Johns Hopkins University and Fast Company at the Hopkins Bloomberg Center, leaders in tech, energy, and academia explored the way AI is changing the way we travel, build, and entertain.

AI use in transportation

Nearly 40,000 Americans died in traffic accidents in 2024. Self-driving cars could help reduce these deaths, according to the autonomous vehicle company Waymo, which operates its self-driving cars in five U.S. cities.

Waymo uses AI in perception systems that help the car navigate the road, prediction systems that help it anticipate other drivers, and planning systems that help it chart the best path forward, according to Smitha Shyam, Waymo’s senior director of engineering. Additionally, Waymo uses generative AI and advanced world models to create simulations that help teach autonomous vehicles in addition to their real-world training on public roads.

Shyam noted that independent assessments have found that “the Waymo driver is better than average human driving on a variety of safety dimensions,” including fewer airbag deployments, injury-causing crashes, and insurance claims. The newsletter Understanding AI and tech news site Ars Technica have supported this claim after analyzing Waymo crash data.

AI’s effect on energy

Electricity demand is continuing to grow, and high power bills are a core issue for the Trump administration and voters, as utility costs were a prominent theme in November’s elections. While many factors can drive up electricity costs—inflation, natural gas prices, climate change, and data centers—and not all states see the same impacts, government forecasters expect retail electricity prices to continue increasing faster than inflation through 2026.

The energy demand from AI is real, though how much that demand will increase and how soon is still unknown, said Calvin Butler, president and CEO of Exelon.

“What we do know: It’s a lot [of demand],” Butler said. “It’s more than what we have, and we have inadequate supply right now to meet that demand, which is why we’re dealing with this issue.”

Keeping up with demand and addressing affordability will require “an all-of-the-above approach” that includes wind, solar, gas, and nuclear, Butler added. “All electrons are good electrons at this point if we’re going to focus on the affordability issue,” he said.

AI for new material development

AI is also affecting material science, helping researchers accelerate the discovery and testing of new materials for extreme conditions, such as hypersonics or atmospheric entry.

Designing for these specific conditions requires combining many different types of existing materials—which generates many potential design options—or creating new materials entirely, which are expensive to make.

“This is where AI has really made a difference,” said K.T. Ramesh, Alonzo G. Decker Jr. Professor of Science and Engineering at Johns Hopkins, in an interview with the Hopkins Bloomberg Center. “It allows you to rapidly look at the material space—a space of possible materials—and figure out what you need to make.”

AI can then help rapidly test potential materials and analyze the results to guide experts toward the best choice for a specific application.

“Both on the fundamental [research] side and on the applied side, these AI methods allow us to speed up the process of doing the work,” Ramesh added. “It allows you to think about problems differently, because you’re no longer worried about ‘Can I do it?’ You’re thinking more about ‘What do I want to do?’”

AI for sports analytics

Through research at the Johns Hopkins Sports Analytics Research Group, AI is also stepping onto the playing field to enhance sports analytics in partnership with multiple sports teams and leagues.

Using AI, researchers are able to mine sports analytics for new insights that impact in-game strategy, particularly in baseball, since it’s a sport that progresses over a series of distinct actions. This information is being used to develop players and protect them from injury, according to Anton Dahbura, co-director of the Johns Hopkins Institute for Assured Autonomy and founder of the Sports Analytics Research Group. Eventually, it can even impact the fan experience through real-time game information.

“We’re really just scratching the surface so far with how we can use AI in sports across the board,” Dahbura told the Hopkins Bloomberg Center, noting that the Sports Analytics Research Group is starting to work on AI analysis for sports with a more continuous flow, such as soccer and basketball. And the research can have broader downstream effects, too, Dahbura said.

“There are so many opportunities in other areas for applying what we learn in sports analytics,” he said. “Things like medical devices and wearables, also the optimization techniques, data engineering.”