Legacy Application Modernization for Utilities: How an AI-Augmented Approach Reduces an Unseen Risk
Your legacy utility systems contain decades of operational logic embedded in configurations, integrations and custom extensions.
AI has officially entered the classroom. According to a 2025 Microsoft report, student use of AI for schoolwork jumped 26% in a single year, with students using it primarily to brainstorm, summarize and generate feedback. I know firsthand the extent kids are using AI as a shortcut in school — one of my kids is a college student and sees almost every student use it heavily, whether it is allowed or not.
Far more than another classroom tool, AI is reshaping how an entire generation learns. While students and employees are rapidly integrating AI into their routines, far less attention is being paid to how it is shaping their thinking and whether it’s making learning easier or just producing shallower insights. The risk is not that students will use AI, but rather that they will rely on it too much before they learn to think independently.
“AI is not like going from a chalkboard to a whiteboard or from cursive to typing,” says Highlights for Children CEO Kent Johnson. “It’s fundamentally different, and we don’t yet understand its long-term impact.”
At the same time, there is work being done to help kids learn to appropriately use and interact with AI. This has implications beyond individual students and their families, because what happens in education with AI now will shape the future workforce. In fact, some states are already moving aggressively. Texas recently launched one of the first statewide AI-focused school programs, designed to integrate AI literacy and tools directly into the classroom experience.
We need to make sure the next generation of employees is prepared to think critically, lead effectively and adapt in a fast-evolving, AI-powered world.
Human Development Doesn’t Scale Like Technology
There’s a common belief that education hasn’t evolved enough, and that technology can help achieve better outcomes, Johnson says. But the data shows otherwise.
Over the past decade, billions have been invested in education technology. Yet measures of student performance have not improved. For example, according to the National Assessment Governing Board, U.S. reading scores declined again in 2024 for fourth and eighth graders, with no state showing improvement in eighth-grade reading.
While AI will transform many aspects of education, learning itself remains a deeply human process shaped by relationships with teachers, parents and caregivers. It takes far more than access to information and the latest tech tools.
“You’re not going to fundamentally change how a child learns to read or bonds with an adult,” Johnson says. “It’s just human biology.”
This dynamic doesn’t stop at the classroom door. It extends directly into the workplace. Organizations are rapidly deploying AI tools, often without rethinking how employees develop judgment, collaborate or build expertise over time.
When AI shortcuts too much of the learning process, it can create a workforce capable of producing outputs, but less prepared to lead, adapt or think independently. And if students rely on AI before developing foundational skills in school, those gaps only get worse when they enter the workforce.
Most learning environments are still in the early stages of figuring out how to use AI in ways that build capability, not just efficiency. If AI is to be a positive force in education, the key questions are:
- Are we strengthening critical thinking, or bypassing it?
- Are we enhancing learning or simply accelerating output?
Where AI Can Actually Help
None of this is an argument against AI in education. The potential is real, if we are intentional about how we use the technology to strengthen, rather than replace, the human elements of learning.
Historically, educators and publishers have created static resources to help adults support children. AI introduces the possibility of more dynamic, personalized guidance for those adults. AI can evolve many aspects of education, such as how content is delivered, how learning is supported and how instruction is personalized.
Now, with AI, it’s possible for companies like Highlights to create dynamic, personalized support for adults, Johnson says. “There is a ton we could do to allow a parent or caregiver to customize their experience. Or say a family gets an issue of our magazine—AI could provide a guide on what the parents need to know to have a great experience at bedtime reading this with their child. Things like that could be improved with AI but still maintain their authenticity and human touch. It’s not all or nothing.”
AI is also beginning to deliver something education has long struggled to achieve—personalized learning at scale. Tools like Khan Academy’s Khanmigo and platforms like Duolingo and Century Tech adapt in real time, identify gaps and provide on-demand support tailored to individual students.
A parallel effort is emerging to ensure AI capability is matched with AI literacy. Digital literacy can’t just be about using tools. It has to also include instruction on judgment, behavior and values, skills that will carry directly into how individuals operate as employees and leaders.
Programs like Highlights’ “Be Internet Awesome,” developed in partnership with Google and distributed to over a million kids for free, offer a glimpse of what this can look like in practice. The initiative focuses on principles like being smart, alert, strong, kind and brave online.
When brought to life through hands-on activities, classroom kits, and community-based programs, these concepts become more than guidelines. They become shared experiences between kids and adults, whether in schools, libraries, or at home.
Importantly, resources like “Be Internet Awesome” are designed to be accessible and flexible. They don’t require extensive training or rigid implementation. They meet educators and caregivers where they are, acknowledging the reality that not every learning environment fits into a formal curriculum model.
Programs developed with the MIT Media Lab, along with curricula from Common Sense Media and initiatives like those from the Connecticut State Department of Education, are focused on teaching students how to question AI outputs, recognize bias and avoid overreliance.
The challenge goes beyond AI adoption. We need to balance AI’s capabilities with the judgment required to use it well.
A More Deliberate Path Forward
We’ve already seen what happens when technology outpaces our understanding of its impact on young people. Social media was, in many ways, a large-scale uncontrolled experiment. It was adopted rapidly, with little foresight into its effects on young people’s mental health, identity and social development. The data we now have suggests the consequences have been significant and are still unfolding.
AI risks following a similar path. “We can’t experiment with AI in education on a generation of kids, because they only get one chance,” Johnson says.
Children don’t get a reset on their development. The choices made today will shape not only how they learn, but how they think, work and lead in the future.
The goal isn’t to slow AI adoption. It’s to approach it with greater intentionality by focusing not just on what AI can do, but on how it shapes human development. The long-term impact of AI in education won’t be defined by access to the technology. It will be defined by whether we use it to strengthen thinking or replace it.
The question is no longer whether AI belongs in education or the workplace. Students are already using the technology, and there’s no going back. The real challenge is whether we will design systems that develop the judgment to use it effectively. If we don’t, we risk building a generation optimized for output, but underprepared to think, adapt and lead.
