AI's Real Impact Isn't in Silicon Valley - It's in Hospitals and Schools
Journalist Josh Tyrangiel, former deputy managing editor at Time magazine, argues that artificial intelligence's greatest potential lies not with tech companies, but with doctors, educators, and government officials who use it to solve concrete problems.
Tyrangiel's new book, "AI for Good," examines where AI actually changes lives. He found that answer by moving past two familiar narratives: AI will cure cancer, or AI will destroy humanity.
"There's something in between," Tyrangiel said in an interview. "What's this actually good for? What's going to change people's lives?"
A Hospital's Sepsis Solution
At Cleveland Clinic, one of the nation's top healthcare systems, administrators made a critical decision: doctors, not technologists, would determine how AI gets deployed.
The clinic deployed AI software to predict sepsis, a deadly immune response to infection that kills roughly 350,000 Americans annually. The system monitors patient signals and alerts physicians when sepsis may be developing, allowing treatment with antibiotics before the condition becomes life-threatening.
The results: sepsis mortality dropped 41% over one year. The clinic saved 1,000 lives.
"AI is not perfect, but we saved 1,000 lives - that seems pretty good," Tyrangiel said, paraphrasing the clinic's leadership.
This kind of work is happening in AI for Healthcare settings across the country, often without the media attention given to tech startups.
Educators Reshape How AI Tutors
In education, Khan Academy and OpenAI partnered to build Khanmigo, an AI tutor. The challenge proved harder than expected.
GPT was trained as an assistant, not a teacher. Building something that actually tutors - that asks questions instead of providing answers - required rethinking how the technology works.
Students treated it as another app. Teachers, particularly veterans in Indiana, found it transformative. They converted decades-old lectures into interactive labs and used the tool to identify struggling students in real time.
"It was a really useful tool for teaching and grabbing engagement and making kids more social and changing learning," Tyrangiel said.
Skepticism Remains Essential
Tyrangiel emphasized that healthy skepticism toward AI is not only understandable - it's necessary.
Young people have watched social media companies overpromise and underdeliver for 20 years. That wariness, he said, will force AI companies to deliver tools that serve the public rather than simply maximize profit.
"We're going to have to demand that the companies deliver these uses the way we want them," Tyrangiel said. "If not, we're just going to sit downstream of whatever they give us, whatever is most profitable, and we're going to suffer for it."
The Window Is Narrowing
Tyrangiel warned that the opportunity to shape AI's direction is closing. Despite predictions that improvement would plateau, the technology continues to accelerate.
"We only have so much time in which we can mold it," he said. "This is going to take skepticism and involvement and real passion if we want it to work the way we want it to work."
For government workers, the stakes are direct. AI for Government applications are already being tested in public agencies. Understanding how these tools work - and where they fail - matters for anyone in public service.
Your membership also unlocks:
