How Nonprofits Are Using Machine Learning To Support Those in Need

Just a decade ago, machine learning was only accessible to the largest, most well-funded companies and research institutions. Opportunities for nonprofits and nongovernmental organizations (NGOs) to use machine learning to advance their work were simply out of reach due to high costs and lack of technical expertise. That has all changed with the rise of cloud computing.

Machine learning is now more accessible to organizations of all sizes and levels of technical expertise because of new cloud-based services that make it easy to incorporate intelligence into applications and operational processes. As machine learning has become mainstream, nonprofits and NGOs are deploying this sophisticated technology into their platforms, solutions and products to better serve their beneficiaries, increase their impact and take their mission to the next level.

Here are three key takeaways from nonprofits that are using machine learning to support those in need, including the homeless, patients suffering from fatal diseases, refugees, human trafficking victims and other vulnerable communities.

Connecting People to the Resources They Need, When They Need Them

Time is of the essence when people in need turn to a nonprofit for assistance. If individuals don’t get the right resources in a timely manner, the results can be detrimental to their well-being. Machine learning has a unique role to play in these scenarios. One great machine learning use case is the ability to match people quickly with information and resources that can make a meaningful difference in their lives — oftentimes much faster and more efficiently than a human could.

For example, Breastcancer.org — a leading digital resource for people affected by breast cancer — uses machine learning to personalize the patient education journey. Receiving a breast cancer diagnosis can be confusing and overwhelming, which often yields an internet search of complex medical terms and difficult to navigate resources. To mitigate those challenges, Breastcancer.org is building a tool that will personalize the path to finding expert information and support which will help individuals make informed decisions about their breast cancer treatment. Patients simply create a profile in a private and secure personal account on the website. Then, Breastcancer.org will use advanced machine learning technology to provide custom article recommendations for individuals navigating treatment and diagnostics. This solution helps patients find credible information to make important choices for their health based on their unique cancer situation.

As another example, the Los Angeles-based nonprofit PATH is using machine learning to address homelessness. For someone living on the street or in a homeless shelter in Los Angeles, the wait to get housing through government programs can take several months. Timing in these situations is often critical; a person who is ready to come in and get help one day may not return the next. To shorten this wait time, PATH developed LeaseUp, a platform that connects clients in real-time with the best possible housing for their needs. The platform uses machine learning technology to capture relevant information about available units of housing in the area, so that case managers can recommend the best housing option to their clients in real time. With this technology, PATH has been able to match more than 850 individuals experiencing homelessness with housing — and reduce the time it takes to find accommodation from 90 days to 45 days.

Scaling the Impact of Nonprofit Staff and Humanitarian Workers

According to data from the U.S. Census Bureau, the median employment of a nonprofit is about four employees, and more than 99% of nonprofits have fewer than 500 employees. Emerging technologies, like machine learning, can have an outsized impact on what nonprofits are able to accomplish with limited staff, resources and funding.

Polaris, a 150-person nonprofit, operates the U.S. National Human Trafficking Hotline, which provides 24/7 support 365 days a year for survivors of human trafficking to get connected to help and stay safe. Over the past 13 years, Polaris reports that calls have increased by over 955% and those directly from victims and survivors have increased by over 4,000%. To help manage this massive influx, the nonprofit integrated a voice bot powered by machine learning to direct non-urgent general information requests to informational resources. Now, more staff time is available to engage with calls from victims and survivors with urgent needs. Within the first six months after rollout, Polaris automated over 1,700 general information calls using the voice bot.

The nonprofit Tarjimly is leveraging technology to scale and create an outsized impact by using machine learning to improve the lives of refugees and the efficiency of humanitarian services. About 75 million people are currently displaced worldwide and these refugees are challenged with language barriers every day. Tarjimly built a tool to eliminate language barriers in times of crisis, enabling volunteer multilinguals to match with refugees who speak their language and dialect in real-time. When a refugee, asylum seeker or humanitarian worker requests a translator for a particular language pairing, Tarjimly’s machine learning matching algorithm reaches out to the best volunteers available. The first volunteer to accept the request is then connected in a live chat with the person in need where they can text, send voice notes or get on an internet call. To date, more than 20,000 Tarjimly volunteers have helped more than 23,000 refugees and aid workers in critical events.

Identifying Those Most At-Risk

A key benefit of machine learning is its ability to quickly scan and analyze complex imagery and massive sets of images — much more accurately and at a much larger scale than humans ever could.

Orbis, an international nonprofit that brings people together to fight avoidable blindness, developed a tool powered by machine learning that can detect common eye diseases, such as diabetic retinopathy, in seconds. It does so by examining digital photographs of the back of the eye. This is game-changing for increasing access to early detection, which is critical to prevent treatable eye conditions from leading to vision loss.

Machine learning is also being used to fight child sex trafficking online and help recover child victims. The nonprofit Thorn uses machine learning to power its Spotlight tool, which enables investigators to more efficiently and quickly sift through thousands of escort ads where child sex trafficking victims are advertised. Approximately 200,000 of these ads are posted each day on the internet. Spotlight’s machine learning models analyze new ads in real time and use intelligent image analysis and natural-language processing (NLP) to flag those that match risk profiles for child victims developed in cooperation with law enforcement agencies.

By leveraging machine learning, nonprofits can stay focused on their mission, while accelerating their pace of innovation and delivering impact in new ways. When it comes to machine learning, we’ve just woken up and haven’t even had our first cup of coffee yet. There is still so much unrealized potential for machine learning and how it can materially benefit society by empowering those organizations charged with a mission to help our most vulnerable communities. And I’m incredibly optimistic about how machine learning can help solve some of the world’s toughest challenges.