For more than a decade, the idea of “Tech for Good” has centered on a simple belief: governments and nonprofits can improve social outcomes by leveraging the expertise and tools of technology companies. Traditionally, this has meant tech firms offering discounted software, engineering support, or infrastructure to public institutions.
There have been meaningful wins. When HealthCare.gov struggled during its 2013 launch, a rapid “tech surge” brought in private-sector experts to stabilize the system. Similarly, organizations like GiveDirectly have partnered with Google.org to deploy tools like mapping and machine learning for faster disaster response.
Yet today, this model is under pressure—and the reason is scale.
When Reach Becomes the Problem
Public-interest technologies often struggle to match the massive reach of consumer platforms. In education, for example, purpose-built AI tools can significantly improve learning outcomes. A structured AI tutor in Ghana has shown results equivalent to a full year of schooling—at a fraction of the cost.
But while such tools are still scaling, billions of users are already engaging with general-purpose AI embedded in platforms like Meta Platforms’s ecosystem—WhatsApp, Instagram, and Facebook. These tools are quick, accessible, and often used by students to generate instant answers.
The downside? Learning can suffer. Studies show that while students may perform better with AI assistance, their understanding declines once the tool is removed. Without guidance, these systems often shortcut the learning process, weakening critical thinking and long-term retention.
The Rise of “Cognitive Shortcuts”
Emerging research highlights the concept of “cognitive debt”—where over-reliance on AI reduces mental engagement. Instead of working through problems, students outsource thinking to machines, impacting memory and comprehension.
In many cases, these AI tools operate without the support systems that make learning effective—like teachers, structured feedback, or curriculum alignment. The result is convenience without depth.
Data reflects this shift. In countries like India, a majority of students using edtech rely on messaging apps rather than specialized learning platforms. Similar patterns are emerging globally, raising concerns among educators and researchers.
Not Just a Technology Problem
The issue isn’t just about AI itself—it’s about where and how it’s used. Platforms optimized for engagement and advertising naturally prioritize ease and speed over thoughtful interaction. Their scale and network effects make them hard to replace.
This creates a new reality: learning is no longer confined to classrooms or textbooks. It’s increasingly shaped by AI tools sitting in students’ pockets.
Public institutions have faced similar challenges before. During the COVID-19 pandemic, health agencies struggled to keep up with misinformation spreading on social media. With AI, the risks may be even more immediate and personalized.
Rethinking the Path Forward
If AI can both help and harm, the real question becomes: which impact scales faster?
To address this imbalance, a broader approach to “Tech for Good” is needed:
1. Reverse the Flow of Innovation
Instead of only transferring technology to public institutions, there’s an opportunity to embed public-sector knowledge into widely used platforms. Tech companies like OpenAI and Google are already experimenting with “learn modes” that integrate educational best practices into AI systems.
2. Build Better Benchmarks
Standardized evaluation tools—such as OpenAI’s health-focused benchmarks—can help measure how well AI systems serve public needs. Expanding these benchmarks across sectors like education, agriculture, and language inclusion could drive better outcomes.
3. Strengthen Safeguards
Parental controls and age-appropriate settings are a start, but they may not go far enough. Imagine AI systems that don’t just block harmful content but actively promote healthy behaviors—like guiding students through problem-solving instead of giving answers outright.
4. Consider Smart Regulation
While some companies have introduced safeguards voluntarily, others act only after harm becomes visible. Governments may need to step in—but carefully. Overregulation risks politicizing technology, while underregulation leaves users exposed.
5. Invest at Scale
Public-sector innovation is often underfunded. While nonprofits may receive a few million dollars, private tech firms invest billions. Bridging this gap is essential if socially beneficial tools are to compete effectively.
6. Educate the Users
Even with better tools, people need to know how to use them wisely. Schools, parents, and governments must teach individuals how to engage with AI—whether for learning, health, or civic participation.
A Turning Point for Tech and Society
The future of “Tech for Good” depends on adapting to a world where scale determines impact. Beneficial innovations alone are not enough—they must reach people at the same speed and magnitude as potentially harmful ones.
If even part of the promise of AI-driven transformation holds true, then both governments and tech companies will need to rethink their roles. Collaboration must deepen, strategies must evolve, and the definition of impact must expand.
Because in the end, the real risk isn’t just harmful technology—it’s the missed opportunity to build something better at scale.
