Booz Allen is rethinking job taxonomy and education for the AI era

Booz Allen Hamilton’s VP of artificial intelligence knows building a future-ready workforce isn’t just about hiring data scientists and engineers.

“I think everyone is realizing [AI is] going to be an entire workforce question; it’s not just this sub-population of mathematicians and computer engineers and whatnot,” Joe Rohner told Tech Brew.

Today, the firm has more than 4,000 employees in its AI and data practices, Rohner said. But those roles are changing.

“What I’m seeing evolve—and we’re doing this within our own job taxonomy now—is we’re acknowledging that there’s a technical side to AI but there’s actually a much larger side on AI workforce, AI adoption, AI consultant, for lack of a better term,” Rohner said, pointing to a set of employees focused on adoption, responsible AI, and change management.

“Those two populations have to work together,” he said. “They both have to understand each other because AI is not just a mathematical model that we run in a protected environment—it’s going to be something that’s out there.”

Building the foundation: Rohner said Booz Allen is approaching firm-wide AI upskilling in a few ways, including through an “AI Ready” training program and through the firm’s internal “badging program,” which offers recognitions like an “AI expert” badge.

But despite upskilling and hiring AI talent “as fast as we can,” Rohner said. “I still don’t have enough AI talent.”

That’s where the firm’s nonprofit partner, the AI Education Project (aiEDU), which is focused on creating “equitable learning experiences that build foundational AI literacy,” comes in.

When it comes to how and when to insert AI into educational systems, there’s a lack of consensus, and, ironically, a lack of data, aiEDU CEO Alex Kotran said. But the answers to those questions are critical.

“In a world where we have tools that are basically going to supplement the need for content knowledge, and broadly speaking, technical knowledge, what are going to be the differentiating skills that students are going to need?” Kotran said, pointing to a need to highlight both computational thinking and “softer skills” like creative and critical thinking and interpersonal skills.

“One of the things we don’t know is, when is it the right time to introduce a language model to a student in a way that doesn’t actually stunt their development?” Kotran added. “So if you give a kid, an 8-year-old or even a 5-year-old, a language model, is that actually going to reduce their ability to become a good writer, to develop writing skills? The answer could be yes. The answer could actually be that it's too much, it’s too powerful for us, and it's something that needs to be introduced at the right time in the same way that, if you’re teaching arithmetic, you don’t just get out calculators and then have them learn to use the calculator—they need to learn the fundamentals of math.”

There’s a “really big need” for additional research on those topics, as well as coordination between companies, researchers, and educators, according to Kotran, who spoke to Tech Brew from Los Angeles, where he was meeting with the LA Unified School District.

“Zooming out, the more important core skill is this orientation toward lifelong learning,” Kotran said. “In the context of AI, it’s really a license to constantly explore and try to understand new tools, and figure out how they can be applied to whatever domain kids are interested in.”