- Tecnologia e Sociedade
- Impacto Social da Tecnologia
- Ética na Tecnologia
- Acessibilidade
- Engajamento Cívico
- Privacidade de Dados
Technology with Social Purpose
Whoever Defines the Problem Owns the Solution
For my TCC, my undergraduate capstone project, my group and I built a system to connect animal rescue NGOs with people interested in adopting. The pilot worked. And the project stopped right there, because a system needs maintainers, and none of us managed to keep it going. That was the first time I understood that technology with social purpose depends on more than good code. In this post I separate corporate "tech for good" from impact that actually changes something, lay out four concrete paths for a developer, share what I learned by doing, and end in the place almost nobody looks: your own job.
Tech for Good or Social Technology?
Every big tech company now has a social impact page. "AI for good" programs, annual sustainability reports, hackathons with nonprofits. Some of it helps someone. But a lot of it is marketing, and the difference is not always in the intentions of the people doing it.
Researcher Ben Green summed up the problem in a 2019 paper: computing's efforts toward "social good" lack a solid grounding in politics or in theories of social change. They rely on vague notions of what is "good" and move forward through small reforms, always with technology at the center. Nobody asks who decided that this was the problem.
In Brazil, that question has a tradition of its own. Renato Dagnino, a professor at Unicamp, uses the term social technology for technology developed with a community and controlled by it, not simply delivered to it. He criticizes conventional technology precisely because it takes control over the work process away from the worker:
"[Conventional technology], by not allowing the direct producer to control the work process, always makes necessary a boss, a capitalist, a manager, a foreman, or, in more modern times, an engineer."
Look at the end of that sentence. In conventional technology, the engineer, and today the developer, is the one who decides on behalf of the people who will use it. That is why I use two tests to tell impact apart from marketing.
- Does it survive without the company? Real impact continues when the sponsorship ends. Marketing disappears along with the campaign.
- Whoever decides, controls. If the company defines the problem, it also keeps the solution, the data and the credit. If the community does not decide, it does not control.
Both tests apply to companies, but they apply to us too. A volunteer developer who shows up at a nonprofit with a ready-made solution repeats the same mistake on a smaller scale.
Four Ways to Use What You Know
If your question is "where do I start?", these are the four paths I recommend. None of them requires quitting your job.
1. Websites and systems the nonprofit can maintain
According to the TIC Organizações Sem Fins Lucrativos 2025 survey, Cetic.br's study of Brazilian nonprofits, only 46% of nonprofit organizations in Brazil have a website or web page. Among those that do, 10% had their site built by volunteers, and volunteers run the social media accounts in 29% of them. There is room, and a lot of it.
But the criterion is not what you know how to build, it is what the nonprofit can maintain after you leave. A simple site, in a tool the team can edit on its own, is worth more than a custom system that only you understand.
2. Accessibility
The WebAIM Million 2026 report analyzed the home pages of the 1 million most visited websites in the world: 95.9% had automatically detectable accessibility failures, with an average of 56.1 errors per page. Portuguese-language pages fare worse: 61.1 errors on average, and the .br domain reaches 62.5. Six basic error types, such as low contrast, images without alt text and form fields without labels, account for 96% of everything.
Brazil has 14.4 million people with disabilities, according to the 2022 Census, and the share reaches 27.5% among people aged 70 or older. Brazil's Inclusion Law (Lei Brasileira de Inclusão, or LBI) requires company and government websites to be accessible. Accessibility is not a favor: it is the law, and it is your audience.
3. Teaching technology
Almost every developer underestimates the value of teaching the basics. TIC Domicílios 2025, Cetic.br's household internet survey, shows that only 54% of people aged 60 or older use the internet, compared with 85% of the general population. Of the 28 million Brazilians who do not use the internet, 16 million are 60 or older. Before any app, there is an access barrier, and it will not be solved by yet another app.
4. Free software, open data and civic tech
There are public-interest projects in Brazil that accept contributions from developers today:
- Brasil.IO: makes Brazilian public data accessible. The project's manifesto says that "restricting access to public data is making democracy elitist". During the pandemic, a task force of 40 volunteers compiled the bulletins from the state health departments.
- Querido Diário: an Open Knowledge Brasil project that makes municipal official gazettes searchable. Its documentation has a dedicated track for people who want to start contributing.
- Software Público Brasileiro: the Brazilian public administration's free software catalog, with 81 systems.
What these projects have in common is that they do not belong to a company. They do not depend on sponsorship to exist, but they do depend on people to keep going, and that is where you come in.
What I Learned by Doing
Before technology, my school of community organizing was Rotaract Club Barueri, where I was Director of Public Image in the 2019-2020 term, along with projects like planting more than a thousand trees in the Baixada Santista region. Then came the projects with code: the animal rescue NGO capstone at Fatec de Barueri, free redesigns of websites for nonprofits and social movements, and basic computer classes for older adults. Those projects left me with four lessons.
The problem was not just technical. Process, time and people weighed more than the system. The code was the easy part.
Without someone to maintain it, it dies. The capstone got the top grade and a pilot that worked. Even so, it did not continue: a system needs maintainers, and at the time none of us managed to keep it going. Today I know that maintenance is not a detail of the project: it is the project.
I learned more than I taught. Listening to the people who would use it changed the way I build software to this day. What seemed obvious to me was rarely what the person on the other side needed.
Access is the first barrier. In the classes for older adults, it became clear that, before any app, what is missing is knowing how to use it and having access. No digital solution reaches people who have not yet reached the digital world.
The Traps for People Who Want to Help
Good intentions protect no one from these four traps. I have fallen into a few of them.
A project with no maintainer
This was my capstone's mistake, and it is not just a student mistake. Software Público Brasileiro, created in 2007, started having technical problems in 2009, and maintaining the portal became harder and harder until it was rebuilt by UnB and USP between 2014 and 2016. If even a federal government project degrades without someone to look after it, a system built over a weekend for a nonprofit degrades even faster. It is the same problem I discuss in The Developer's Manifesto: infrastructure everyone uses and almost nobody sustains.
Solutionism
Evgeny Morozov coined the term solutionism for the temptation to treat political and social problems as if they were technical questions of efficiency. A donation app does not fix the lack of public policy for abandoned animals. A scheduling system does not fix the shortage of doctors. Technology can help within a problem, but it is rarely the answer to it.
Exploited unpaid labor
Carlos Montaño, a professor at UFRJ, shows how the discourse of the "third sector" helped justify the state's withdrawal from areas like social security. The critique is not aimed at the people who work in NGOs, who often do what nobody else will. It is aimed at the model in which volunteers cover, for free, what should be public policy or paid work. When a for-profit company asks for "volunteers" to build its social project, be suspicious.
Doing "for" instead of "with"
Dagnino argues that technology should be produced with the participation of the people who will use it, and calls the idea of bringing science "to" citizens naive. In practice: before writing a single line of code, ask what the organization needs, what it already uses, and who will look after whatever you deliver. If you cannot answer that last question, it is not yet time to start.
The Impact That Fits in Your Job
The place where a developer has the most social impact is not the weekend project. It is the everyday work, which reaches thousands or millions of people. Four decisions fit there.
- Ship accessible work even when nobody asks for it. Contrast, alt text and form labels do not need anyone's approval. WebAIM's six most common errors are in the hands of whoever writes the front end.
- Collect only what is necessary. The LGPD, Brazil's data protection law, has a principle called necessity: limiting data processing to the minimum required for its purpose. When someone asks for "one more field, just in case", that principle is your argument.
- Refuse dark patterns. The FTC, the US consumer protection agency, defines dark patterns as design practices that trick or manipulate users into making choices they would not otherwise make and that may cause them harm. Hidden cancellation, pre-checked consent, fake urgency. Saying no to that is part of the job.
- Collective organization. One person alone can rarely push back on a product decision. An organized team can. The decisions above only hold up when they do not depend on one developer's individual courage. I wrote about collective ways of organizing tech work in Platform Cooperatives and the Fediverse.
This is where the two tests from the beginning come back. The impact you build at your job survives your departure if it becomes the team's standard. And whoever decides, controls: if developers take part in the decision, it stops belonging to the company alone.
Conclusion: Start Small, Start Together
Technology with social purpose is not a product category or a page in an annual report. It is a question about power: who defines the problem, who controls the solution, and who keeps looking after it when the enthusiasm fades. My capstone taught me this in the simplest way, with a pilot that worked and a project that stopped because no one was there to maintain it.
If you want to start, do not start by building yet another app on your own. Pick one front:
- Offer help to a nonprofit in your city, and first ask what it already uses and who will maintain things.
- Fix the most common accessibility errors in the product you work on.
- Teach a basic technology class at a community center or to someone in your family.
- Open the Querido Diário or Brasil.IO repository and look for an issue to get started.
Start small, and start together with a real community. A project other people can carry on makes more of a difference than the best-built system nobody can maintain.
References
- Dagnino, Renato. Tecnologia Social: contribuições conceituais e metodológicas. EDUEPB/SciELO Books (2014).
- Green, Ben. "'Good' isn't good enough". NeurIPS AI for Social Good Workshop (2019).
- Morozov, Evgeny. To Save Everything, Click Here: The Folly of Technological Solutionism. PublicAffairs.
- Montaño, Carlos. "O projeto neoliberal de resposta à 'questão social' e a funcionalidade do 'terceiro setor'". Lutas Sociais (PUC-SP).
- Cetic.br. TIC Organizações Sem Fins Lucrativos 2025, table C1.
- Cetic.br. TIC Organizações Sem Fins Lucrativos 2025, table C14.
- Cetic.br. TIC Domicílios 2025, table C2.
- Cetic.br. TIC Domicílios 2025: principais resultados (key findings).
- WebAIM. "The WebAIM Million: The 2026 report".
- IBGE. "Censo 2022: Brasil tem 14,4 milhões de pessoas com deficiência" (2025).
- Brazil. Law 13.146/2015 (Lei Brasileira de Inclusão, Brazilian Inclusion Law).
- Brazil. Law 13.709/2018 (LGPD, Brazilian General Data Protection Law).
- Federal Trade Commission. "Bringing Dark Patterns to Light" (2022).
- Brasil.IO. Manifesto.
- Open Knowledge Brasil. Querido Diário (documentation).
- Governo Digital. Software Público Brasileiro.
- UnB/USP team. Technical history of the rebuild of the Software Público portal (2017).