a young woman looks at her phone in class

This School Year, Make AI Work for Learning

BlogArtificial IntelligenceSep 30, 2025

As students attend school this fall, artificial intelligence (AI) will be a constant presence in their lives. Tools like ChatGPT and Gemini are now part of students’ daily routines, used to draft essays, solve math problems, and summarize complex readings. But despite its widespread use, most schools still lack clear policies or guidance to ensure that AI tools support learning rather than undermine it.

Recent headlines highlight growing concerns about AI in education. Nearly half of students are using generative AI for schoolwork, often without clear guidance from teachers or school systems. A new study finds that students who relied on AI to write essays did not perform better than those who wrote independently. Media outlets are raising alarms about students losing their ability to think critically and educators are reporting an uptick in passive learning behaviors. Further, recent reports indicate that, in some cases, AI chatbots exacerbated students’ mental health risks—for example, by encouraging self-harm or intensifying psychosis—highlighting potential threats to their well-being. These trends are unfolding alongside survey data showing that most schools are not teaching students how to use AI ethically.

The increasing use of AI underscores a broader gap in how schools equip students to navigate emerging technologies responsibly so that their learning and well-being are supported. Without intentional instruction, students may use AI in ways that shortchange learning and compromise their well-being.

Schools are at a crossroads: They can either view AI as a risk to manage or approach its integration with care and responsibility, enhancing its potential to support both academic growth and student well-being. At Child Trends, we support the latter. Over the past two years, we have worked alongside school leaders, teachers, and researchers to develop three research-based frameworks that guide responsible AI integrations in school systems: the AI Coherence Framework, the AI-Class Framework, and the AI Risk Framework. These tools help schools move from uncertainty to action so that AI strengthens teaching, builds student agency, and protects the safety and well-being of all learners.

A Clear Path Forward: Coherence, Capacity, and Caution

1Coherence: Align AI with what matters most in schools.

AI use in schools should be guided by coherence, meaning alignment with instructional goals, curricular scope and sequence, and pedagogical best practices. Our AI Coherence Framework outlines four levels of alignment:

  • Technological: Is the tool safe, transparent, and age-appropriate?
  • Curricular: Does it support what students are learning, when they are learning it?
  • Pedagogical: Does it reinforce critical thinking, creativity, and active engagement?
  • Implementation: Is it usable and supported in classrooms?

When AI tools are misaligned with district goals and daily instruction, they risk undermining student learning, widening differences in access to resources, and fueling mistrust.

2Capacity: Build schoolwide readiness for AI implementation.

Even the best-aligned AI tools fail in the absence of informed users: That’s why we created the AI-Class Framework, a guide for school leaders and teachers to implement AI responsibly and effectively within schools and classrooms. The framework includes four components:

  • Awareness: Understand current comfort and experience levels with AI.
  • Knowledge: Build educators’ skills in prompt engineering, data privacy, and instructional integration.
  • Collaboration: Engage teachers, students, and school leaders in co-designing practices and policies.
  • Alignment: Ensure that new AI tools fit into, rather than disrupt, existing curricula and routines.

Used together, these components support implementation coherence and help schools turn AI from a disruptive force into a thoughtful tool for innovation.

3Caution: Manage the risks of AI in schools.

AI is not neutral. Algorithms can reinforce biases, jeopardize privacy, widen digital divides, and harm student mental health. Our AI Risk Framework helps education leaders assess and mitigate risk across eight domains, including fairness, transparency, community engagement, and the risk of inaction. As we outline in our recommendations on AI regulation, effective oversight must go beyond compliance. It should ensure that AI systems are aligned with educational values and protect student well-being, and that they’re shaped with input from those most affected by their use.

For example, schools that avoid AI entirely may unintentionally leave students behind, particularly in communities already facing digital opportunity gaps. However, introducing AI without strong safeguards or design processes that include teachers and students can also deepen harm.

The takeaway is clear: Risk is not a reason to avoid AI, but school leaders should address it strategically, ethically, and in partnership with the communities that schools serve.

Leading Into the Future

This school year, education leaders have an opportunity to guide AI integration into schools with coherence, care, and courage. This includes teaching students how to use AI thoughtfully and critically, supporting teachers as they adapt to new tools, and ensuring that AI strengthens the core mission of education: preparing young people to think, create, and thrive.

At Child Trends, we support schools and districts at every step of AI integration. Our research-based and field-tested frameworks provide a clear roadmap to help school systems leverage AI in ways that reflect their values, promote instructional quality, and support the learning of all students.

Suggested citation

Holquist, S.E., Kelley, C., Haugen, M., Aceves, L., & Li, W. (2025). This school year, make AI work for learning. Child Trends. DOI: 10.56417/956k2277u

Webinar on Responsibly Integrating AI into Education

From AI Literacy to AI Coherence: How Educators Can Integrate AI Responsibly in Teaching and Learning
Tuesday, October 21 | 2:30–3:30pm EST | Register here