Growing up, Justin Steele was "a sensitive, brainy kid" who spent a lot of time thinking about what he could do to improve people's lives. After earning an engineering degree from the University of Virginia, he received a master's in urban social policy and nonprofit management at Harvard and went to work in the nonprofit sector full-time. Since 2014, he has held senior positions with Google.org, where he's taken a lead role in the organization's work on inclusion, education, and economic opportunity.
PND recently spoke with Steele about Google.org, its efforts to develop AI tools for nonprofits, and what it is doing to address homelessness in the Bay Area.
Philanthropy News Digest: What is Google.org, and how much does it award annually to nonprofits here in the United States and globally?
Justin Steele: Google.org is Google's philanthropic and charitable arm. We support nonprofits that are working to address challenging problems and try to apply scalable data-driven innovations in support of those efforts. What's unique about Google.org is that we were established when the company went public with a commitment of 1 percent of its equity and an ongoing commitment of 1 percent of its net profit for charity. Google.org is the biggest beneficiary of that 1 percent ongoing net-profit commitment, and we currently award more than $300 million in cash grants to nonprofits globally each year, roughly split 50/50 between the U.S. and internationally.
PND: Can any nonprofit apply for a grant?
JS: We are predominantly invite-only in our philanthropy, but we do have a model called the Impact Challenge where we invite nonprofits to participate by sending us their ideas. Sometimes the challenge is topic-based, sometimes it's based on geography.
In the U.S., we are currently running Impact Challenges in a number of geographies. We have a $10 million Impact Challenge open in the Bay Area and $1 million challenges open in Georgia, Minnesota, Nebraska, and Ohio. A panel of local experts who have influence in the states where the challenge is occurring help us narrow down the candidates. The panel chooses the finalists who receive funding, but we also open it up to a public vote. The People's Choice winners get extra funding at the end.
The state-level Impact Challenges change from year to year, although this is the third time we've run a challenge in the Bay Area, which is where we’re headquartered. Last year, we ran challenges in Illinois, Nevada, and Colorado, and we expect to launch new challenges in other states in 2020.
We also opened up the AI Impact Challenge globally in 2018 and 2019 for organizations that are working on interesting applications of artificial intelligence for social good.
PND: You and your colleagues recently issued a paper titled "Accelerating Social Good With Artificial Intelligence." What does AI have to do with social good? And can you give us an example of an AI-driven project that has delivered real results in terms of social good?
JS: AI is an emerging field. It's obviously transforming a lot of Google's tools, and we use it internally every day, which is why we call ourselves an AI-first company. But Google.org also wanted to make sure we're thinking through how AI can be used to solve some of our biggest social challenges. The AI Impact Challenge generated over twenty-six hundred applications from a hundred and nineteen countries, lots of them interesting, especially in the areas of the environment, conservation, and energy. If nothing else, it showed us that there are a lot of nonprofits and social enterprises eager to use AI tools in their work.
Google itself has used AI in its flood-forecasting tools, with interesting results, especially in places like India, which is prone to devastating and deadly flooding. We take historical data and look at areas that have been impacted by flooding in the past. Then we use real-time river-level data, take Google maps data with topography and elevation profiles, pump that into the models, and run thousands of simulations. Based on that information, we're able to issue alerts to warn people where and when flooding might occur.
We've also seen a number of AI-driven applications submitted through Impact Challenges focused on similar issues such as rainforest health and irrigation in Africa. One of our grantees — Thorn, a nonprofit in the U.S. that works to stop the trafficking of children — has seen particularly strong results. We gave them a $2.5 million grant and sent a team of five Googlers to work with them for six months to improve their models. The result was a spotlight tool that uses clustering algorithms to find victims faster. Every year, more than two hundred thousand escort ads are posted daily in the U.S. across different platforms. Because traffickers often move their victims across state lines, the same ads tend to pop up across different geographies. The algorithms we were able to develop with Thorn examine information in posts such as phone numbers, location, linguistic style, and image similarities. Using that information, law enforcement can triangulate the location of a trafficked person. Today, more than seven thousand law enforcement agencies in all fifty states and Canada are using the tool and have identified and located more than twenty-eight thousand trafficked individuals.
We're still in the early days, and the technology is developing rapidly. That's partly why we launched the Impact Challenges. We're curious and eager to help develop and train more nonprofits to use technology to advance their missions.
PND: Your background is in engineering. Has that training been helpful in terms of your role at Google?
JS: Google's obviously an engineering company, so it's definitely helpful to have that background. On a practical level, engineering helped me lose my fear of complexity. In college, I studied chemical engineering and did insanely difficult problems. That experience taught me how to break a problem into its component pieces and get down to first principles. You learn that you can break any problem into its component pieces and build back to the solution. I took a lot of that approach with me when I went to grad school and then when I started to work for nonprofits trying to understand the factors driving complex social problems.
For example, in criminal justice reform, where we've awarded $40 million over the last five years, we took a data-driven approach to finding solutions. We initially invested in a broad theory of change focused on advocacy and job creation and other things. But we ended up putting most of our resources into efforts to understand, from a data-science perspective, what's happening in policing, what's happening in sentencing, what's happening in jails and prisons.
For instance, the Vera Institute of Justice recently published a major report on rural incarceration. We gave them $5 million and sent an engineering team to help them build data tools that would help them understand what's happening in jails at the county level across the country. There are thousands of counties in the U.S., and it's really hard to get a handle, especially in real time, on what's happening there. One hypothesis was that incarceration was growing faster in rural counties than in urban areas, but it was unproven. With the resources and engineers we sent them, Vera was able to demonstrate that incarceration in rural areas is in fact increasing at a very significant rate while urban incarceration is actually decreasing. You see cities like New York trying to close its Rikers Island facility, but in rural areas the number of people who are being jailed is spiking. The opioid crisis is a significant driver of the increase. And the way our carceral system works, a lot of people end up being held in jail because they can't make bail. Parole violations are another significant driver of why people are incarcerated in rural areas. As I say, it's an emerging picture. But Vera's report is driving a lot of policy makers to think about what we, as a society, can do to address some of the challenges we're seeing in these areas.
PND: Last spring, Google.org awarded grants totaling $50 million to address homelessness and displacement in the Bay Area, while your parent company committed $1 billion to build affordable housing in the region — with other tech giants announcing similar pledges. Are efforts by private employers sufficient to solve the Bay Area's homelessness and affordable housing crises, or does a real solution lie elsewhere?
JS: I don't think there's a solution to housing in the Bay Area that's independent of the public sector. We see our role as a philanthropic entity supporting innovative ideas that the public sector might not be able to support in the short term. But if you want to solve it at the systemic level, it really requires partnerships with the public sector.
One example I can give is our work with Hamilton Families, a nonprofit based in San Francisco that works with homeless families. Back in 2014 they had an idea to partner with the San Francisco Unified School District to identify students who were experiencing homelessness or at risk of becoming homeless. Obviously, homelessness is really disruptive for kids who are trying to succeed in school. So SFUSD decided to create a rapid response team to prevent families from losing their homes and rehouse them when they did. We invested $1 million in that project, which helped lead to a 40 percent reduction in the average wait list for family shelters in the city. The City of San Francisco subsequently invested $4.5 million in the organization's Project Every Home, which was an outgrowth of its work with SFUSD. And there was an additional $20 million of private capital that eventually flowed in as well.
We know we're not going to solve this alone. But we do have flexible philanthropic resources we can use to take risks and look for things that might take longer to generate results. Plus, we can invest in innovation, create proof points to demonstrate outcomes, and then have the public sector and other private investors come in and fund some of those ideas.
As you noted, Google itself has committed $1 billion to address the problem, and those resources will help build twenty thousand additional housing units in the Bay Area. But we need more than twenty thousand. Again, that's why partnering with the public sector is critical.
As far as developing a more diverse workforce here at Google, I'm not sure we can go anywhere but up. I'm optimistic we'll have a more diverse workforce in ten years, but it's certainly not going to happen on its own.
One of the initiatives we announced a couple months ago with SFUSD Superintendent of Schools Tony Thurman was a $10 million rising STEM scholars initiative. With our partners, we did some research and found there were three thousand low-income students of color in the Bay Area who were qualified and capable of succeeding in advanced placement STEM courses but who were not currently enrolled in those courses. That's a huge missed opportunity for Google and other technology companies in the Bay Area. And that’s why we’re investing $10 million in five nonprofits that are working to identify those students, work with the schools directly to get those students into AP classes, support those classes with resources to make sure the schools are prepared to teach that content, and, maybe most importantly, invest in the teachers who teach those classes. A lot of times, students are interested in taking AP courses, but especially in the case of AP computer science, we find a dearth of qualified teachers who are credentialed to teach the subject at the high school level. In the first year, we piloted the initiative at sixteen schools, and we've already doubled the number of black and Latinx students in those schools taking AP STEM classes. So the next step is to scale that across the whole Bay Area.
So we're investing significantly in initiatives that are designed to strengthen the pipeline of local talent into places like Google. It's tough, and it's really tough to do it at scale. It's going to take a concerted effort. And it's going to take more than just Google. But we're confident that, working with others, including the public sector, we can figure it out and disrupt the status quo.
— Matt Sinclair