AI for Good: How Dataminr is Transforming Humanitarian Assistance

In 2025, an estimated 305 million people will require urgent humanitarian assistance and protection, according to the United Nations. Of those 305 million, 190 million people are facing life-threatening needs. At the same time, funding for the organizations that address these issues is declining, especially as financial support from the United States Agency for International Development (USAID), the largest donor of aid worldwide, remains in limbo after the Donald Trump administration froze it in January. 

As a result, humanitarian organizations have to plan ahead and use their resources wisely to reduce the impact of growing crises. Artificial intelligence (AI) is an integral tool, helping organizations analyze large amounts of data, spot early warning signs and respond faster. The AI company Dataminr is making that technology accessible for more organizations. 

“AI is really good at creating efficiencies, freeing up time for higher-order tasks. But at its best, it can help move from reactive to proactive crisis response. Maybe you’re not fully ahead of the crisis before it starts, but you’re able to intervene sooner and understand the cascading impacts,” said Jessie End, vice president of social good at Dataminr. “That was really the ‘holy grail’ vision for our AI for Good program — to not only leverage Dataminr’s expertise and product but also to help humanitarian organizations navigate an increasingly complex crisis ecosystem.” 

The AI for Good program helps social impact organizations and U.N. agencies use AI to improve awareness, predict risks and strengthen humanitarian efforts. Part of the program includes working with Dataminr on custom AI projects.

"We’re continuing to bring in more data sources and elements that can help detect earlier signals, but what we’re really aiming for is something more predictive,” End said. "Our focus — especially through the AI for Good program — is to develop custom AI projects that help organizations stay ahead of crises.”

Transforming crisis response with AI

Two to three nonprofits are selected to create custom products each year through an open proposal process. Over six to nine months, they work with the Dataminr team to develop AI solutions that benefit not only the partner organization but also the wider humanitarian sector.

“For example, one of our 2024 partners, Mnemonic, manages a massive data archive of human rights violations that can be used for international prosecutions. But with so much data — text, audio and video — prosecutors struggle to find what’s relevant,” End said. “Our project with them focuses on automating the tagging process so the data is easier to navigate. And this doesn’t just help Mnemonic, it supports prosecutors worldwide who rely on this information to pursue justice.”

The program also offers organizations free or discounted access to Dataminr’s First Alert tool, which tracks and analyzes critical information for crisis response and management in real time. The tool monitors over a million sources of public data — including social media, government websites, air quality monitors and earthquake sensors — saving organizations time and helping them make informed decisions.

“Our AI platform then processes billions of data points from these sources and is trained to detect events,” End said. “It acts as a sense-making machine, identifying critical developments in real time."

In response to funding shortages in the humanitarian sector, Dataminr also recently launched an initiative that provides nonprofits operating for the public benefit with free access to First Alert.

The importance of collaboration

Working closely with partners in this way and getting their input before launching new tools is key to making something that’s actually useful. 

“I’ve made mistakes in this context, and one example is when we tried to launch a crisis response program that offered free licenses to organizations responding to major disasters,” End said. “It seemed like a great idea. But we quickly realized that for smaller, under-resourced NGOs, learning a new technology in the middle of a crisis was impossible. Now looking back, it’s obvious. But at the time, we hadn’t asked enough questions before launching it. Your partners know their needs better than you do, so involve them from the start.”

It’s an important realization, as ensuring that smaller organizations benefit from this technology can make a big impact. 

"A lot of corporate social impact initiatives focus on big, well-known nonprofits because they bring credibility,” End said. “That’s fine. But if you really want to make an impact, you need to work with smaller, local organizations that most people have never heard of. These are the groups doing the heavy lifting on the ground, often with far fewer resources. Designing solutions that fit their needs — including their limited bandwidth and infrastructure — makes all the difference,”

End also stressed the importance of increasing collaboration in the tech sector, despite the tendency for companies to be protective of their intellectual property.

“When it comes to early warning systems, for example, you need multiple components,” End said. “We do one piece of that. The only way to build truly effective solutions is to work together and stop being so protective of data and knowledge. With crises escalating worldwide, we can’t afford to operate in silos.” 

Measuring success

"Measuring success in our space is really challenging," End said. "A lot of our users are security professionals responsible for keeping staff and beneficiaries safe in high-risk areas. Success often means preventing something bad from happening, but how do you measure something that didn’t occur?”

Because of this, Dataminr relies heavily on anecdotal evidence to evaluate the impact of its tools. Its partnership with the medical aid logistic nonprofit Direct Relief offers one example. 

“Their job is to get medical aid and equipment in crisis situations,” End said. “They were using our First Alert platform during the early days of the invasion of Ukraine to understand needs on the ground. Our platform allowed them to know where and when to send power generators to hospitals in Ukraine.”

Time saved seems to be a common denominator of measuring success, too. “Whether it’s reducing response times in a crisis, or cutting down manual work like tagging attacks on human rights defenders, our goal is to free up staff for higher-priority tasks and help organizations act faster," End said.