Amira Learning is designed to be equally accessible to girls and boys, yet we recognize that structural and societal factors can influence how learners of different genders experience and benefit from the product.

In our research and field observations, we have noted that girls often outperform boys in early literacy measures, which can sometimes lead to under-identification of boys who need support. Amira helps address this disparity through objective, real-time assessment and AI-driven scaffolds that flag reading difficulties based on performance—not teacher perception—ensuring that all students receive targeted support regardless of gender.

In certain contexts, especially in low-income or rural households, gendered responsibilities and mobility restrictions may limit girls’ access to digital tools. To mitigate this, Amira is optimized for short, flexible use sessions and can be accessed on low-cost, shared devices—enabling learning during limited windows of availability and without requiring high levels of adult supervision. Its low-bandwidth functionality further supports use in regions with unstable connectivity, where access may be constrained by household priorities or infrastructure.

We also acknowledge that girls and women often carry greater caregiving responsibilities, which can limit time for education or engagement with their children’s learning. Amira addresses this by including simple, automated progress reports and home access options, allowing caregivers—regardless of literacy level or time constraints—to support reading development with minimal additional burden.

By collecting and analyzing sex-disaggregated usage data, Amira can continuously monitor for engagement gaps and tailor training, support, and product improvements to promote gender-equitable outcomes.

In sum, while Amira is designed for universal usability, we remain attentive to the structural gender inequalities that can shape learner access and outcomes, and we actively design features and implementation strategies to reduce these barriers.

Amira Learning is purposefully designed to reduce systemic barriers to literacy for marginalized and disadvantaged children. The platform supports learners from diverse backgrounds—particularly those underserved by traditional instructional models—through inclusive content, accessible technology, and adaptive, real-time support.

For children with limited access to quality instruction, Amira acts as an always-available reading tutor, listening as they read aloud and delivering individualized feedback and scaffolds without requiring adult supervision. This provides critical practice and support, even in environments where one-on-one instruction is not feasible.

Amira is optimized for low-cost devices and low-bandwidth environments, enabling use in under-resourced schools, rural communities, and low-income households. Its bilingual capabilities in English and Spanish support emergent bilinguals and migrant learners, while its speech recognition models are trained on diverse accents and dialects to fairly evaluate oral reading across populations.

For children with disabilities, Amira includes accessibility features such as screen reader compatibility, audio narration, and simplified navigation. The platform’s scaffolded support—ranging from decoding hints to comprehension prompts—helps students progress at their own pace, reducing frustration and supporting confidence.

Amira also includes culturally responsive stories that reflect a broad range of identities and experiences, allowing children from ethnic minority groups, refugee communities, or non-dominant cultures to see themselves in the learning process.

By combining inclusive pedagogy, equitable AI, and adaptive design, Amira ensures that all learners—especially those historically marginalized—have access to the tools they need to become confident, fluent readers.

Amira Learning is purposefully designed to uphold UNICEF’s core equity and inclusion principles, ensuring all children and especially the most marginalized, have access to quality literacy instruction, regardless of their background, ability, or learning environment.

1. Representation and Culturally Responsive Content
Amira’s reading library includes over 4,000 leveled texts that reflect diverse characters, family structures, and global cultures. Content is regularly reviewed to ensure it is free from bias, culturally respectful, and inclusive of learners from underrepresented backgrounds. Stories can be adapted to reflect local values and norms in international contexts.

2. Accessibility and Universal Design
The platform meets WCAG 2.1 Level AA standards, offering screen reader support, audio instructions, and simplified navigation. It is built to serve children with disabilities, emergent readers, and those with limited digital literacy. Amira is optimized to run on low-cost devices and in low-bandwidth settings, making it accessible for children in rural, low-income, or humanitarian contexts.

3. Multilingual and Local Adaptability
Amira is currently available in English and Spanish, with bilingual voice guidance and customizable content pathways to support language learners. Pilots in Colombia, Mexico, and the UAE have included adaptation of text, narration, and illustrations to better reflect local languages and cultures.

4. Fairness and Bias Mitigation in AI Design
Amira’s speech recognition models are evaluated across a wide range of accents, dialects, and reading profiles to ensure equitable accuracy and minimize bias. The AI provides consistent feedback across learner groups and avoids punitive labeling or categorization. Educators can review and override AI outputs to retain instructional control.

5. Pricing Equity and Access
Amira offers context-sensitive pricing, including donor-supported deployments, PPP-based discounts, and free caregiver access via school-licensed models. This reduces economic barriers and supports public education systems in delivering foundational literacy at scale.

6. Inclusive Product Development
The design process integrates feedback from diverse user groups—including students from low-income schools, English learners, and children with special needs. The internal team represents a mix of cultural, linguistic, and socioeconomic backgrounds, which helps inform inclusive decision-making and design.

Amira Learning embodies UNICEF’s digital inclusion and equity objectives by proactively designing for marginalized learners, mitigating digital and instructional gaps, and ensuring that no child is excluded from the opportunity to learn to read.

Amira Learning is built with a strong commitment to AI equity, ensuring that all students, regardless of background, language, or ability, receive fair, transparent, and effective support in their literacy journey.

To mitigate bias, Amira’s speech recognition models have been trained and validated on diverse datasets that include a wide range of accents, dialects, reading abilities, and first languages. Extensive testing has been conducted with English learners, early readers, and students from historically marginalized groups to ensure consistent accuracy and avoid demographic skew in error detection or fluency scoring.

Amira avoids opaque decision-making by keeping AI output transparent and human-reviewable. All reading assessments generate clear, traceable scoring and feedback, which teachers can audit, override, or adjust. This maintains educator agency and supports accountability in instructional decisions.

In addition, Amira’s design avoids free-form AI generation during instructional moments. Instead, it delivers scripted, research-backed scaffolds to ensure instructional integrity and appropriateness for young learners. Privacy and data safeguards are also in place to uphold child rights and protect user data from misuse.

Through equitable AI design, validation, and oversight, Amira promotes fairness, accessibility, and learning success for all students.