Thursday, April 22, 2021


AI for good: Tech company and nonprofits protect forests, track river pollution and deliver clean water

Veronica Combs 
 TechRepublic
EARTH DAY 2021
 Image: GettyImages/Christopher Kimmel
© Image: GettyImages/Christopher Kimmel

Volunteering for a day to help a nonprofit with a tech project is nice. Teaching the organization's team members to build and manage their own tech solutions in the long term is sustainable. That's the approach DataRobot took with its AI for Good program.


The artificial intelligence company started the program in July 2019 after working with the Global Water Challenge in 2018. DataRobot team members analyzed more than 500,000 data points to predict future waterpoint breaks in rural communities in Africa.

Chandler McCann, general manager and data science practice lead for the public sector and head of AI for Good at DataRobot, said the company wanted to offer nonprofits more than a one-time event like a hackathon or a difficult-to-maintain custom-code solution. The goal of the program is to help organizations create meaningful AI systems that are easy to use and self-sustaining.

"After formalizing the program, we partnered with GlobalGiving to tap their extensive history and expertise in working with networks of NGOs and nonprofits to help us facilitate the design of the application process and vet applicants," he said.

The first cohort in the AI for Good program worked on these two projects:
Entel Ocean: The digital unit of Entel, a telco in Chile and Peru, developed IoT sensors to detect and predict when forest fires are likely to occur. The data collected by these devices is paired with DataRobot's platform, allowing the organization to detect a fire 12 minutes earlier than the previous monitoring systems.
Anacostia Riverkeeper: The group built a system to predict E. coli levels in the waterways in one of the first times AI has been used to predict water quality across the whole Chesapeake Bay region.

McCann said each solution was developed collaboratively between the team at DataRobot and the non-governmental organizations and nonprofits.

"DataRobot subject matter experts work closely with AI for Good beneficiaries to evaluate the problem they're trying to solve, ensure it's appropriate for AI, and then begin to train employees and build models to work toward a solution," he said. "Subject matter experts are always on hand to help, guide, and refine the process of getting the models up and running, as well as how the organization can optimize results and expand the deployment."

DataRobot team members also train the teams who will use the platform to ensure they have the skills they need to maintain the project over time. McCann said the program has been a real proof point that AI can have a positive impact on the world.

"These are use cases that we as a company may have never experienced if we didn't create this program to help nonprofits and NGOs," he said.

Nonprofits apply to participate in the program and applications are open now for the second cohort. The application has four sections: Organization summary, project description, implementation and optional supporting information. The deadline to apply is April 30.

The AI for Good team at DataRobot evaluates applications on three dimensions:
Impact: 

How will a proposed project impact a community and the people who live in it?

Framing: Does the project take on a compelling challenge and what data supports the solution?

Readiness: Does the organization have the capacity and support from leadership to make the project successful?

No comments:

Post a Comment