The United States was and still remains the single largest donor of humanitarian aid in the world.

In fiscal year 2023 alone, the country accounted for 43% of all government funding donated to humanitarian causes across the world — including $43.8 billion from federal programs like USAID to help combat global poverty, disease, and food insecurity.  But over the last year, the Trump administration terminated 83% of USAID programs and laid off most of the agency’s workforce. USAID officially closed its doors in July 2025. 

The cuts are projected to cause more than 14 million additional deaths worldwide by 2032, including 4.5 million deaths among children under five, according to a peer-reviewed study published in the global health journal The Lancet.

As the United States scales back investments in international development and social good, the future remains uncertain for global aid programs and the millions of people who depend on them. 

Keeping a global activist mindset, students in Development Engineering programs at UC Berkeley are charting a path forward for their own field. 

In April, DevEng 210 students Jeremy Lowe, Hila Mor, Mariam Ibrahim, Maksymilian Jasiak, and Diana Bolanos presented their analyses of past and prospective funding pathways for a field that traditionally receives a large part of its funding through government investment. 

“We’ve been asking some important questions: ‘How can our field continue to grow? How can research survive funding cuts?’’ Jasiak said.    

“It was important that this conversation be grounded in reality — so we looked at just how Development Engineering projects could pursue various alternative funding streams,” he added. 

Their work examined a field based on development practices that have been refined over decades to adapt to global challenges. Beginning in the 1940s, early conceptions of the “Third World” and the perceived shortcomings of modernization inspired researchers to design small-scale technologies for underserved communities. 

By the 1990s, critiques of globalization had pushed the field toward community-centered solutions that could be scaled locally, sustained financially, and developed through interdisciplinary collaboration.

Today, Development Engineering combines engineering, economics, and social science approaches to address the world’s most pressing humanitarian issues — ranging from global poverty and water insecurity to public health and disaster response.

However, developing these solutions and bringing them to scale requires funding. 

Students examined a range of historic funding sources for development projects, including public and private investment, philanthropy, and foreign assistance from organizations such as USAID, UKAID, the World Bank, and United Nations agencies. Their findings suggest that many of those traditional funding streams are drying up. 

“Going forward, there will be a much higher bar for what gets funded and what is pursued,” Lowe said. 

Students identified local governments as one potential source of sustainable funding through tax revenue. Yet many developing countries face significant obstacles in raising the funds needed to support development programs independently. Citing research from the United Nations Conference on Trade and Development, the group noted that developing countries lose an estimated $100 billion annually through tax avoidance. 

With billions of dollars in foreign aid disappearing, philanthropists and private donors have stepped in to fill some of the sudden funding gaps. Students pointed to Project Resource Optimization, which mobilized $110 million in private funding for approximately 80 former USAID-funded initiatives serving an estimated 41 million people worldwide.

Still, they concluded that philanthropy may help keep some critical projects alive in the short term, but cannot replace international aid at scale. Instead, the future of development may depend on hybrid funding models that combine public investment with revenue-generating strategies capable of attracting private capital. 

As traditional aid dollars shrink, Development Engineers are now tasked with both developing technologies that improve lives, and building the financial systems needed to sustain them. 

Yet funding is only one of the transformations shaping the future of the field. 

As artificial intelligence becomes increasingly embedded in efforts to fight poverty, distribute aid, and guide public policy, another group of DevEng 210 students turned their attention to a different question: How can these technologies be used without compromising privacy, equity, and human rights?

Sophie Pesek, Daniel Bostwick, Jasmine Hughley, and Susana Constenla-Villoslada presented research on the risks of AI-powered development projects, exploring how data is collected, analyzed, and ultimately used to make decisions that affect vulnerable communities.

AI is increasingly used in international development to identify communities in need, predict poverty levels, and target aid distribution. By analyzing mobile phone metadata, satellite imagery, and other large datasets, organizations can gather information that once required expensive surveys and years of fieldwork. 

However, students argued that the same systems capable of expanding access to information can also create new kinds of risks. 

Their research found that many development-focused AI systems rely on passively collected data, including phone records, mobility patterns, and satellite imagery, often without the knowledge or explicit consent of the people being analyzed.

According to the students, privacy risks emerge long before an algorithm makes a decision. Even when data sets are anonymized, mobile phone records and location data can act as proxies for sensitive information, revealing details about a person’s health, socioeconomic status, and background. 

“AI systems treat people as numbers or stacks of data before they’re human beings, so there’s a big issue with algorithmic targeting,” Bostwick said.

As a result, populations may be categorized, monitored, or targeted without meaningful control over how their data is used. However, the risks do not end once data is collected.

Students also pointed to concerns about the labor behind AI systems used in development, noting that data-labeling work is often outsourced by AI companies to workers in the Global South who are paid relatively low wages despite helping to build technologies that primarily benefit wealthier countries.

“At the end of the day, we are a capitalist society that tries to find legal loopholes, and this is a great example of finding loopholes to get the best possible AI model,” Bostwick said. 

According to their research, many development organizations now use machine learning systems to determine who receives aid, healthcare services, or other forms of support. While these tools can improve efficiency, mistakes in algorithmic decision-making can carry significant consequences, potentially excluding individuals from critical services or directing limited resources away from those who need them most.

AI systems also have the potential to produce group-level harms by reinforcing stereotypes, stigmatizing communities, or systematically excluding entire populations based on flawed assumptions embedded in the data.

“The biggest issue facing data privacy is really how the algorithm’s decisions will affect people,” Bostwick said. “These are real problems — they’re not abstract.” 

Rather than treating privacy as a single technical problem, students advocated for a “privacy-by-design” approach that incorporates safeguards throughout every stage of a system’s development and use.

Their recommendations included stronger transparency requirements, differential privacy techniques, greater community participation in system design, and policy regulations that give affected populations more control over how their data is used.

“AI for development is not just a technical challenge, but a governance and social justice issue that has real impacts on people across the world,” Bostwick said.

As artificial intelligence becomes increasingly embedded in efforts to fight poverty and deliver aid, students argued that a major challenge will be to ensure technologies designed to help vulnerable communities do not inadvertently expose them to new forms of surveillance, exclusion, or inequality.

“Data privacy is not a single-stage issue,” Bostwick said.