The Human Touch: Balancing AI and Agency in Education

Integrating emerging technologies into any educational setting requires a human touch to ensure they are used appropriately and responsibly. Because technology can sometimes widen divides, the need for a human-centered approach is perhaps most pronounced in learning environments that have traditionally suffered from a lack of resources and include learners that have been systemically pushed to the margins of the education system. A human-centered approach to artificial intelligence (AI) can enhance the benefits of AI—increasing learning opportunities, providing rapid feedback, and personalizing learning—while mitigating the challenges of bias, inaccuracy, and inaccessibility.

This International Day of Education, we’re spotlighting the importance of human-centered AI. By equipping educators and learners with AI literacy and embedding human skills, oversight, and decision-making into AI-enhanced programming, we ensure that technology use is driven by human capabilities and experiences. Here’s how we’re making this vision a reality at World Education.

The Spectrum of Human-AI Collaboration

Using AI in education is not always a yes-or-no decision. Instead, integrating AI usually falls along a continuum of human-AI collaboration. On one end of the continuum, an AI-enhanced tool takes on a more autonomous role, managing tasks like data gathering, analysis, and even decision-making. For example, an AI-enhanced tool might guide learners through personalized lessons, assess their performance, and determine progression with minimal human involvement. On the other end, AI provides a support function, assisting educators with tasks like generating ideas, analyzing trends, or providing accessible resources, while humans remain fully in control of the learning experience. The most effective implementations lie somewhere in between, where AI augments human efforts, streamlining processes or offering insights to educators and learners.

Human-Centered AI in Learning Interventions

Pace AI’s Teacher-Centered Features

The team at Pace AI is working with World Education to develop and scale their tool that supports English language learning for adults in the U.S. Emphasizing the importance of a human-in-the-loop, the Pace team built the tool to enable high levels of teacher influence. While Pace’s AI automatically generates exercises and vocabulary lists, which teachers can edit and decide which exercises are administered to learners. In this way, Pace combines the personalizing power of AI with teachers’ knowledge and experience. Pace continues to prioritize teacher and learner feedback as they build new features and make enhancements to the tool.

Set Up for Success with Teacher and Parent Involvement

In a similar field testing project in Ghana, World Ed researchers evaluated the use of Google Read Along as a tool for providing literacy programming to out-of-school girls in Ghana. Read Along uses AI speech recognition to offer targeted feedback on the girls’ reading, providing tailored support to build language skills. The researchers found that substantial involvement of Mentor Teachers and parents/caregivers to guide lessons was critical to the success of the pilot. While the AI tool was engaging and showed great promise, the intervention as a whole needed humans to set up the structures (e.g., providing tablets, overseeing daily 15-minute sessions) and monitor and evaluate progress, bridging the gap between technology and learners.

Supporting Digital Skills

For community case workers in Malawi who deliver health information, Adobe Liquid Mode improves access to training materials by using AI to reformat documents for mobile devices. This makes critical content accessible to users with limited internet bandwidth or outdated devices. These AI-driven accessibility features are powerful, but case workers need support from people to understand how to make the features work well for them. For most, adopting any new tool requires additional or increased digital literacy. By providing flexible, human-centered guidance and training, we maximize the tool’s impact.

Our work doesn’t stop at integrating AI tools. We focus on providing individuals with the knowledge and skills to use these tools effectively and ethically. Through initiatives like Digital Resilience in the American Workforce and the GenAI Maker Space, we are developing frameworks and training to help educators and learners build AI competencies. The CampGPT Prompt Book equips teachers with strategies to develop student-centered prompts for generative AI, emphasizing the balance between automation and human judgment. The AI Integration Framework provides questions for educators and decision-makers to consider their own capacity to effectively implement AI-enabled tools.

As AI continues to reshape education globally, a human-centered approach is more important than ever to ensure ethical and effective outcomes. This International Education Day, we invite you to join us in advocating for human-centered AI literacy for everyone.