Editorial Note
This article is intended for educational and informational purposes. It summarizes research-based commentary published by the Brookings Institution’s Center for Universal Education on July 9, 2026. The publication is an expert analysis drawing on existing research and evidence rather than a newly released randomized experiment or clinical trial. New To Education is not affiliated with Brookings or the authors.
Artificial intelligence is frequently presented as a solution to some of education’s most persistent challenges.
AI-powered platforms may provide personalized tutoring, help teachers prepare lessons, translate learning materials, automate administrative tasks, and expand educational support in places where qualified teachers and resources are limited.
However, a new analysis published by the Brookings Institution on July 9, 2026, warns that educational AI could reproduce many of the inequalities it is supposed to reduce.
The authors argue that learners and communities in the Global South must be involved in designing, governing, and evaluating AI-supported education. Otherwise, countries may receive tools developed elsewhere without having meaningful influence over the languages, knowledge, values, and assumptions built into them.
The central message is uncomfortable but important: distributing AI tools is not the same as creating educational opportunity.
What Was Published on July 9
Brookings published “Feminist and Global South Perspectives on AI-Supported Learning Environments” through its Center for Universal Education.
The analysis was written by Bhawana Shrestha, Anthony Luvanda, and Jamila Razzaq. The authors examine how artificial intelligence in education interacts with gender inequality, limited infrastructure, unequal internet access, language exclusion, local knowledge, and institutional readiness.
They identify four major tensions that education leaders should consider:
Access compared with meaningful participation.
Contextual relevance compared with the marginalization of local knowledge.
Gender inclusion compared with the possibility of amplifying existing bias.
Rapid adoption compared with the actual readiness of schools and communities.
The researchers do not argue that schools should reject artificial intelligence entirely. Instead, they call for education systems to examine who designs the technology, who controls it, whose knowledge it represents, and who can use it meaningfully.
Access to Technology Is Not the Same as Participation
One of the publication’s most important arguments is that access figures can hide deeper inequalities.
A student may technically have access to a phone, computer, or internet connection but still lack the freedom, time, privacy, electricity, data, training, or family permission needed to use it consistently.
This can be especially important for girls and young women in communities where access to technology is shaped by gender expectations.
A household may own one device, but control over that device may not be shared equally. A girl may be permitted to use it only at certain times or for limited purposes. She may also have more household responsibilities than male family members, leaving her with less time for digital learning.
Under those conditions, simply reporting that the household is connected does not provide an accurate picture of educational opportunity.
Schools and governments therefore need to measure meaningful participation, not merely device distribution.
That includes examining how often students can use technology, whether they can study safely, whether they have the necessary digital skills, and whether the tools are actually improving learning.
AI Systems Often Prioritize Dominant Languages
AI tools are generally more capable in languages for which large amounts of digital material and training data are available.
This creates a disadvantage for communities whose languages are underrepresented online.
A student may receive a detailed, accurate explanation in a globally dominant language while receiving a less reliable or less culturally appropriate response in a local language.
That difference can affect more than convenience.
Language influences how students understand identity, history, community, science, and culture. When an education system treats only a small number of languages as technologically valuable, it may gradually push local languages and knowledge traditions to the margins.
The Brookings analysis warns that AI systems can reinforce global hierarchies of knowledge by treating dominant perspectives as universal while overlooking local experience.
An educational tool may appear neutral, but its answers are shaped by the information used to create and train it.
Educators should therefore ask not only whether an AI system produces fluent responses, but also whose knowledge those responses represent.
Local Knowledge Could Be Lost in Standardized AI Systems
Many educational AI systems are designed in wealthy countries and later introduced into communities with different histories, school structures, languages, teaching traditions, and social needs.
A tool that works well in one setting may not transfer easily to another.
For example, an AI platform may assume that every student has reliable broadband, individual device access, consistent electricity, small class sizes, or teachers with extensive technical training.
Those assumptions may be unrealistic in rural or under-resourced schools.
The content itself may also reflect unfamiliar examples, cultural expectations, or educational priorities.
When technology is introduced without local adaptation, schools may become dependent on systems that do not fully understand their students.
This does not mean education should reject outside knowledge or international technology. It means that communities should have the authority and resources to adapt technology rather than simply receiving a finished product.
Teachers, students, parents, local researchers, and community organizations should be part of the design process.
AI Could Repeat or Amplify Gender Bias
Artificial intelligence learns patterns from existing data.
When those patterns reflect historical inequality, the resulting systems may repeat it.
An AI career tool, for example, could recommend different occupations to boys and girls based on patterns found in older employment data. A learning platform could interpret student behavior differently depending on biased assumptions built into its models.
Women are also underrepresented in many parts of the technology industry, including the development and governance of artificial intelligence.
That matters because the people building a system influence which problems receive attention and which risks are treated as important.
The Brookings authors argue that women and girls must be included not only as users of educational AI but also as developers, researchers, decision-makers, and community leaders.
A program cannot be considered inclusive merely because girls are permitted to use it.
Meaningful inclusion requires influence over how the technology is designed, tested, evaluated, and governed.
Schools May Be Pressured to Adopt AI Before They Are Ready
Education systems are under growing pressure to demonstrate that they are preparing students for an AI-driven future.
That pressure can encourage schools to purchase technology before establishing clear goals, teacher training, privacy protections, technical support, or methods for evaluating whether the tools improve learning.
The Brookings publication distinguishes between principled concerns and practical limitations.
Some educators may oppose certain uses of AI because they are worried about bias, student agency, cultural relevance, privacy, or the weakening of human relationships.
Others may appear resistant because their schools lack stable electricity, internet access, technical staff, or professional development.
Neither situation should automatically be dismissed as fear of technology.
In some cases, hesitation may be a rational response to a tool that was introduced without sufficient planning.
Responsible adoption begins by asking what educational problem needs to be solved. Only then should leaders decide whether AI is the appropriate response.
Teachers Need More Than Access to New Tools
Teachers are often expected to implement technology after decisions have already been made by governments, school leaders, or vendors.
That approach can leave educators responsible for systems they did not help choose and have not been adequately trained to use.
Successful AI integration requires sustained professional development.
Teachers need opportunities to test tools, identify errors, discuss ethical concerns, redesign assignments, protect student information, and determine when human judgment should take priority.
They also need permission to reject an AI recommendation when it does not fit the student, lesson, or local context.
Artificial intelligence should support professional judgment rather than quietly replacing it.
Teachers understand their students’ backgrounds, personalities, relationships, and changing needs in ways that automated systems may not.
The strongest educational model is therefore unlikely to involve handing instruction over to machines. It will involve teachers making informed decisions about when technology adds value and when it interferes with learning.
Governments Need Local Data and Stronger Governance
The Brookings authors recommend that governments invest in local-language data, community research capacity, digital infrastructure, and systems for auditing AI.
This is essential because countries that depend entirely on imported technology may have limited ability to understand or challenge how those systems operate.
Local researchers should be able to examine whether an AI system produces different results across gender, language, disability, income, or geographic groups.
Governments also need clear rules governing student data.
Education platforms may collect information about academic performance, attendance, behavior, interests, writing, and personal challenges. That information can be highly sensitive.
Schools should understand what data is being collected, where it is stored, how long it is retained, whether it is used to train future models, and whether it can be shared with third parties.
Innovation without accountability can create new risks for the very learners technology is intended to help.
Funding Should Be Connected to Equity Standards
The analysis also argues that organizations funding AI education projects should require evidence of community participation and inclusive design.
A program should not be considered successful because it distributed devices, created user accounts, or introduced an AI platform into a large number of schools.
Funders should ask whether students use the technology meaningfully, whether teachers received training, whether local languages are supported, and whether disadvantaged groups are benefiting.
They should also examine whether communities can continue operating the program after short-term funding ends.
A pilot project may look impressive while outside experts and funding are present but collapse once those supports disappear.
Sustainable educational technology must fit within the financial, technical, and human capacity of the education system using it.
Why This Research Matters Beyond the Global South
The concerns raised by the Brookings authors are not limited to developing countries.
Schools in wealthier nations also face unequal device access, unclear AI policies, biased systems, insufficient teacher training, language barriers, and pressure to adopt technology quickly.
The difference is often one of scale and available resources rather than the complete absence of the problem.
A rural school in the United States, a multilingual classroom in Japan, and an under-resourced school in Kenya may face different circumstances, but all must ask similar questions.
Does the technology fit the learners?
Can teachers understand and challenge it?
Are students’ data and rights protected?
Does the tool strengthen learning, or does it merely create the appearance of modernization?
The July 9 publication offers a useful reminder that educational technology should be evaluated through the experience of the people most likely to be excluded.
A system that works only for students with reliable devices, strong English skills, technical confidence, and extensive support is not genuinely universal.
What Educators Can Learn From the Findings
Educators do not need to become AI engineers to apply the lessons from this research.
They can begin by examining who benefits from a tool and who encounters difficulty.
Teachers should test whether AI-generated materials accurately represent different cultures and communities. They should also check whether translation tools preserve meaning rather than simply producing grammatically fluent sentences.
Students can be taught to question AI output instead of treating it as an unquestionable authority.
That includes asking where information may have come from, what perspectives may be missing, and whether the answer reflects the student’s local environment.
Schools can also involve families and students in technology decisions.
The people using a system often identify practical barriers that administrators and developers overlook.
Key Takeaways
The Brookings Institution published new research-based education commentary on July 9, 2026, examining AI-supported learning through feminist and Global South perspectives.
The authors warn that providing access to AI does not guarantee meaningful participation, especially for girls, rural learners, low-income students, and linguistically marginalized communities.
Many AI systems prioritize dominant languages and knowledge systems, potentially weakening the visibility of local languages, cultures, and educational traditions.
Women and communities in the Global South should be involved as designers, researchers, leaders, and decision-makers rather than treated only as future users.
Schools should not adopt AI without adequate infrastructure, teacher training, privacy protection, local adaptation, and clear educational goals.
Governments and funders should evaluate success through learning, inclusion, community participation, and sustainability rather than device or account totals alone.
The concerns identified in the publication also apply to unequal and multilingual education systems in wealthier countries.
Frequently Asked Questions
What educational research was published on July 9, 2026?
The Brookings Institution’s Center for Universal Education published an analysis titled “Feminist and Global South Perspectives on AI-Supported Learning Environments.”
Was this a peer-reviewed experimental study?
The publication was research-based expert commentary drawing on existing evidence and scholarship. It was not presented as a newly completed randomized or controlled experiment.
What is AI-supported education?
AI-supported education includes student tutoring systems, adaptive learning platforms, automated feedback, teacher planning tools, assessment systems, administrative tools, and technologies that analyze student progress.
Why could AI increase education inequality?
AI may favor students with reliable internet access, personal devices, dominant-language skills, digitally trained teachers, and strong institutional support. Students without those advantages may receive fewer benefits or face new barriers.
Why are local languages important in educational AI?
Students often understand complex ideas best when materials connect with their language and culture. AI systems with weak local-language support may provide less accurate information or marginalize local knowledge.
Does the research say schools should ban AI?
No. The authors argue for more equitable design, stronger local leadership, better infrastructure, gender-responsive policies, community participation, and responsible governance.
What role should teachers have?
Teachers should help select, test, adapt, and evaluate AI tools. Their professional judgment should remain central to decisions about learning and student support.
How can schools evaluate an AI product?
Schools should examine accuracy, privacy, bias, accessibility, language support, teacher control, evidence of learning benefits, technical requirements, long-term cost, and the experiences of students who may be disadvantaged.
Final Thoughts
Artificial intelligence may eventually help education systems overcome barriers involving teacher shortages, limited resources, language access, and personalized support.
But technology does not enter an empty space.
It enters education systems already shaped by inequality, history, culture, gender, geography, and differences in political and economic power.
The Brookings analysis published on July 9 makes the case that those conditions cannot be treated as secondary concerns.
The future of educational AI should not be decided only by the companies building the most powerful systems or the governments able to purchase them first.
Students, teachers, women, local researchers, and communities in the Global South must have real influence over how these technologies are designed and governed.
AI may become a valuable educational tool, but only when access is accompanied by agency.
Without that agency, technology risks modernizing the appearance of education while preserving—or even deepening—the inequalities underneath it.
Related Articles
Japan’s 2026 Generative AI Pilot Schools Could Shape the Future of Classroom Learning
https://www.newtoeducation.com/view-blog/japans-2026-generative-ai-pilot-schools-could-shape-the-future-of-classroom-learning-6a4c66f6944fa
Schools Need Clear AI Rules Before They Buy More AI Tools
https://www.newtoeducation.com/view-blog/schools-need-clear-ai-rules-before-they-buy-more-ai-tools-6a368152834c6
Sources
Brookings Institution — Feminist and Global South Perspectives on AI-Supported Learning Environments
Brookings Institution — Center for Universal Education
Brookings Institution — Global Task Force on AI in Education
