How to tell if AI is actually making an impact

Since I took up AI coding, I’ve built about half a dozen apps. Each solves a specific problem I had, and I use them daily. They’ve given me a tremendous sense of satisfaction. Looking back, what stands out isn’t their complexity or polish, but the impact they’ve had on me.

That got me thinking about what success with AI looks like. We talk a lot about productivity gains and model benchmarks. But for an individual, a company or a country, how do you know AI is actually working? And what counts as impact?

That is the question behind the third and final episode of the Why Memory and Storage Matter podcast with Micron Technology, where we look at what it takes for AI to deliver real impact across industries and communities.

Impact starts with problem solving

For this episode, I sat down with Joshua Lee, corporate vice president and Singapore country manager at Micron Technology, and Koo Seng Meng, director of talent and ecosystem at AI Singapore. Joshua leads Micron Singapore, while Seng Meng spends his days building the national AI talent pipeline.

And the way I hear it, creating impact starts with solving problems. Less than a year into Micron’s partnership with AI Singapore, it has already delivered a series of high-impact, youth-centric and community-focused initiatives.

One such example is the AI For Good Festival, which brought AI learning opportunities to more than 6,000 students. What excited Joshua wasn’t just the number of students, but seeing them work on real problems, whether from industry or from their own lives. It was there that he met a student whose grandparent had fallen at home. The student’s response was a vision system that detects falls and judges when medical assistance is needed.

"AI is more than just technology,” Joshua said. Learning on real problems leads to applications, and ultimately opportunities for the students.

The same logic runs through the fab, where the problem is simple to state and hard to solve. As Koen De Backer explained in the first episode, a single wafer passes through a complex process of more than 1,500 production steps over three months. Keeping that running with as little variation as possible is where AI comes in, says Joshua, with Micron using the data its factory generates to run operations more predictably and with more automation, but always with a human in the loop.

Taken together, these examples point to a simple measure of AI’s value: not how sophisticated the technology is, but what it enables people to do. Whether helping a student address a family concern or improving factory operations, the goal is the same – to make a meaningful difference.

Building an ecosystem takes time

If impact starts with a problem, where do the problems and the people come together? For Seng Meng, that is Singapore’s role. The Lion City’s strength has never been size, he says, but “our ability to connect things together.” Once you have people from different backgrounds in the same place, good surprises happen.

This didn’t happen overnight. AI Singapore was set up in 2017, well before the current wave of interest in AI. Over the years, they have run programmes such as a youth conference and the National AI Student Challenge. These gave students, researchers and companies reasons to meet and work on problems together.

But the stakeholder Seng Meng kept returning to was community. “Everybody who can learn should have the avenue and resources to,” he said. That now means working with NGOs and steering more experimentation towards problems with no obvious commercial return.

Industry has a part here too. Joshua pointed to AI Bridges, Micron’s volunteer-run programme that introduces students to AI fundamentals through hands-on projects, innovation challenges and mentorship. In Singapore, the company is also working with AI Singapore to train educators and volunteers, so that the learning compounds.

Seng Meng also sees this going beyond Singapore. If the ecosystem works well, global problems can be brought here, matched with the right people, and solved. And Singapore is happy to share what it learns with others, he said, because that is how it learns in return.

His yardstick for success? Not how much AI we build, but how AI delivers impact. That is a harder standard than benchmarks, and a better one. It is also, I think, the thread running through all three episodes.

And that brings the series to a close. Across infrastructure, talent and partnerships, one thing remains constant: all resting on memory and storage. Thank you for following along.

Watch the podcast here.