The progress of artificial intelligence is rapid. Within a short period, AI has transformed from an invention in the hands of researchers and tech companies into something we encounter in our daily lives, offices, community services, and development initiatives.
However, as AI continues to evolve, the larger question that is increasingly difficult to ignore is who owns the technology, who is profiting from it, and who is paying the price?
That question was central to Oxfam International’s preparations for Digital@UNGA 2026 and other AI-related meetings at the 81st session of the United Nations General Assembly. In a statement released September 21, Oxfam International said a handful of AI corporations and nations are consolidating massive amounts of power, wealth, data, infrastructure, and knowledge and called for increased global oversight and the involvement of civil society, workers, impacted populations, and Global South states.
Oxfam is looking at how AI could be used to reduce poverty and inequality. AI is a huge opportunity, but we don’t want it to be ‘locked up’ in a handful of powerful companies and countries.
The AI boom is creating a new global inequality debate.
AI is frequently marketed as a technology capable of providing solutions to the world’s greatest problems. It can facilitate scientific and medical research, enhance access to information, aid education, process large datasets, and presumably enable institutions to react more rapidly to humanitarian and environmental disasters.
But these opportunities are not equally available to all.
While anyone can utilize an AI chatbot, this capability is very different from having the data, the computing infrastructure, the expertise, and the capital to build AI machines. Countries with an existing strong tech industry and large capital reserves have a much more advantageous position to develop AI infrastructure and attract more investment.
For countries and communities where access to digital has already been limited, however, the rapid expansion of AI could become yet another form of inequality.
Oxfam says a limited number of companies and countries also now hold much of the data, infrastructure, frontier models, and expertise that power AI. In a briefing, the organization said this poses challenges to understanding who controls AI, who has access to AI, who benefits from AI, and who suffers the fallout from AI.
This brings the AI debate to a whole new level.
It is not always a discourse of power and privileged access.
Oxfam is calling for stronger global AI governance.
A key refrain in the Oxfam statement is that AI governance must go beyond voluntary commitments by tech giants.
It is advocating for governance that’s enforceable, effective, and human rights-based. It states that AI regulation must tackle transparency, accountability, human oversight, human rights due diligence, and people’s access to remedy when affected by AI.
Oxfam also thinks regulation should consider other factors besides safety.
The group says it hopes those making laws will think about the wider impact of AI on inequality, work, participation in society, misinformation, and the environment.
This becomes especially significant due to the fact that AI cannot exist in complete isolation from society. Geography matters in terms of where data is gathered, where the technology is designed, and where infrastructure is created.
Oxfam’s wider campaign on digital inequality also maintains that this can reinforce and even magnify wider inequalities. It estimates there are still billions of people who have no genuine access to the digital age and has warned that marginalized groups may face an even higher risk of harm from new technology while also being further excluded from its benefits.
Why this matters for developing countries
The global AI race is getting worse.
Nations with excellent digital infrastructure, research centers, a workforce capable of performing tasks, and access to significant computing resources can engage in AI to a greater extent. The others might need to depend on innovation from other nations.
That creates a potential long-term dependency.
A country may be able to afford to buy an AI system, but it does not mean that the country can afford to develop, test, adapt, and govern it.
That’s why the AI capacity-building conversation has been taking over.
Developing countries require access to develop technical expertise, research, data capabilities, and digital infrastructure and have a genuine say in the rules of the AI game.
Oxfam believes, though, that the Global South must not just become another market for AI products. Instead, nations and communities must have genuine input into how AI is created and used.
Civil society wants a seat at the table.
Moreover, Oxfam is worried about who will be involved in global discussions on AI.
According to the group, powerful governments and tech companies shouldn’t control the discussions around AI: “Decisions relating to AI cannot and should not be made by powerful governments and corporations alone. Civil society groups, workers, impacted communities, and countries with less influence in global decision-making must also be able to meaningfully participate.
This is an important issue for NGOs.
Grantees working at the community level directly observe the impact of technological change. They may recognize the consequences of a new digital platform for beneficiaries, workers, or vulnerable populations in ways that policymakers, funders, or the providers of technology do not.
An AI system might be technically effective—but for someone with limited digital skills or who speaks a language that does not have much data for the system to learn from, it can lead to issues.
The experience of local NGOs can sometimes be included in the policy debates.
This means their role is not just to use technology but rather to influence the debate on how it should be used.
What does this mean for NGOs?
This debate is particularly pertinent as more non-governmental organizations (NGOs) start to play around with AI.
Nonprofits are leveraging AI for an ever-increasing number of tasks, such as writing proposals, conducting research, analyzing data, creating content, translating, and administrative tasks.
For staff-centric organizations, these tools can potentially free up some time for staff so they can spend more time with programs and communities.
However, it also presents a number of new challenges.
NGOs could use personal data, beneficiary lists, financial data, case records, and so on. If you upload this information to an AI engine, there is a risk of endangering privacy and security.
The issue of human supervision is raised as well.
It’s important to remember that an AI system can, for example, produce a report, summarize something, or detect a pattern in data, but that doesn’t mean the result should always be blindly approved.
Individuals who are aware of the organization and community still have to dig into the decisions and context from which the information was gathered.
For NGOs, responsible AI, then, has to be much more than just taking up new tools.
It must have public-facing external policies, employee training, data protection, and human review.
The digital divide could become an AI divide.
Oxfam’s worries also relate to a wider issue: the digital divide.
Today, people across the globe still lack affordable access to digital technology, while many others have access to ever-more sophisticated AI devices.
Oxfam adds that over two billion people don’t have equitable access to the tools, technology, and platforms of the digital world. It also said that the digital revolution has been highly unequal, with money and power concentrated in the hands of a relatively small number of tech giants and wealthy countries.
This gap will likely grow even larger as AI plays a bigger role in education, jobs, health care, and public services.
One of the most obvious beneficiaries of AI-supported education is the community with good access to the internet.
A non-governmental organization can’t fully utilize data if they lack the proper digital infrastructure.
The lack of computing power may restrict a developing country from developing its own AI ecosystem.
That’s why the question of AI access has to be paired with a larger debate on the development of digital infrastructure.
AI can still be a powerful tool for social good.
Despite its concerns, Oxfam’s position is not that AI should be rejected.
The organization acknowledges that AI has significant potential to help address major global challenges if it is designed, deployed, and controlled responsibly.
For NGOs, that potential is significant.
AI could help organizations reduce repetitive administrative work, process information more quickly, improve communication, identify patterns in program data, and support fundraising activities.
Humanitarian organizations could potentially use AI to analyze large datasets during emergencies. Education organizations could explore AI-supported learning tools. Health-focused NGOs could use data systems to improve planning and outreach.
The important question is how these technologies are introduced.
AI should support the work of people who understand communities rather than automatically replacing human decision-making.
An NGO could use AI to organize information, for example, while staff members make the final decision about what that information means. It could use AI to prepare a first draft of a document while program experts review and adapt it to the local context.
This approach keeps technology in a supporting role.
The environmental side of AI also matters.
It’s not just about data, work, and privacy.
AI infrastructure also has environmental consequences.
The electricity, and sometimes water, consumed by huge data centers can be a lot, too. For increased computing power, questions are arising regarding the impact these will have on the local people and environment.
Environmental impacts are a particular concern, as Oxfam highlights as it responds to the wider debate about governance of AI. Oxfam maintains that the costs and harms of AI shouldn’t be borne by communities who have the least say.
This creates another issue for policymakers.
The issue isn’t just how fast new AI infrastructure can be deployed but also where, how it is supplied with power, and who bears the brunt of its environmental costs.
What should NGOs be thinking about now?
For NGOs, the growing debate around AI provides an opportunity to look beyond the excitement surrounding new tools.
Organizations can start asking practical questions about their own AI use.
- Who is using AI within the organization?
- What information is being shared with AI systems?
- Which decisions require human review?
- Are staff members aware of the limitations of AI-generated information?
- Do beneficiaries know when AI is being used in ways that affect them?
- And does the use of AI actually improve the organization’s work?
These questions can help NGOs introduce technology in a way that supports their mission rather than allowing technology to determine the mission.
This is particularly important for smaller organizations. They do not necessarily need expensive AI systems to benefit from the technology. In many cases, the more important first step is understanding where AI can genuinely solve a problem and where traditional human expertise remains more appropriate.
The bigger conversation is about who controls the future.
Oxfam’s statement at UNGA81 comes at a time when AI governance is becoming an increasingly important international issue.
The organization is calling for global rules that address unequal power within the AI industry and for greater participation by communities and countries that have traditionally had less influence over major technology decisions. It also supports measures that promote open and interoperable systems, digital public goods, and greater AI capacity in Global South countries.
These are Oxfam’s policy positions, but the underlying questions are relevant across the NGO sector.
As AI becomes more deeply integrated into society, organizations will need to think not only about what the technology can do but also about the systems surrounding it.
- Who owns the data?
- Who controls the infrastructure?
- Who makes the rules?
- Who checks the technology?
- And what happens when something goes wrong?
The future of AI should include more voices.
We are still in the early stages of the AI revolution. So many of the governance and regulation that relate to it are in the process of being created.
That makes the current moment particularly important.
Decisions on AI technology made by governments with significant technological power or by companies with vast financial resources leave little room for others to have a say.
NGOs and civil society organizations are able to offer a fresh perspective since they work more closely with the communities experiencing poverty, inequality, marginalization, discrimination, and unequal access to technology.
Their experience can insert problems from the real world into the policy dialogue that might otherwise go on about techno-economic development.
So, increasing civil society influence at the UN is just one aspect of a broader discussion about whom we want to be governing the digital future.
The bigger picture
Artificial intelligence can become a new tool for social development. It can enable actors to operate more quickly, to analyze information, and to find new pathways to reach people. But progress does not necessarily mean sharing the advantages.
The AI debate is becoming a discourse on who owns, is held accountable for, and can access the tech, as much as it’s becoming a debate about innovation.
Oxfam is therefore urging for stronger international regulations and more transparency and participation by civil society, workers, and impacted people. They warn that without this, existing inequalities could be further cemented into the AI economy.
For NGOs, the message is particularly relevant.
AI can allow organizations to be more efficient and effective, but responsible use is not just a matter of using some software and getting to work. Organizations will need to consider privacy, transparency, human oversight, inclusion, and ultimately the communities they seek to impact.
The more significant question, perhaps, is how much.
It almost certainly will.
What does that change look like, and more importantly, who gets to decide?
If AI is to grow further, ensuring that less privileged groups aren’t just part of the audience for its deployment but are actually involved in discussions of its design could end up being one of the biggest challenges for NGOs, governments, and civil society alike in the coming years.
