In an exclusive interview Amen Ra Mashariki, Director of AI and Data Strategies at the Bezos Earth Fund talks about how he works to identify strategies and solutions that will help environmental justice organizations to use data to solve complex issues.
Environmental justice is an intriguing term. What do you encapsulate with that term and what are the projects currently deployed to catalyze environmental justice from the Bezos Earth Foundation?
There are a lot of ways to think about environmental justice, but from a funding perspective, it is about making sure that organizations that are focused on the ground implementation of solutions have an impact on communities that have historically been left out.
One can think about environmental justice as a wider group of organizations focused on a clear set of values to make sure that as we transition to new energy infrastructures and frameworks, no community is left out.
At the Bezos Earth Fund, we have implemented a few things; to enable the implementation of environmental justice, the organizations should have access to both local and detailed data.
We have funded the ability for a couple of environmental justice organizations to come together and work to build out what we refer to as the Environmental Impact Data Collaborative.
The organizations are working together to surface important data, but they are also working with academic institutions to build a platform to house that data.
Can you please tell me about your time at the White House and the work you did there?
I was a White House fellow and then a political appointee during the Obama administration. At the time we weren’t really doing a good job in the federal government in terms of hiring for innovative tech positions.
During my time, data science was the big buzz and we didn’t have a mechanism for identifying and hiring good data scientists into the federal government in order to use data science to solve a lot of public challenges.
My job was essentially to help leadership across the United States federal government to think through the best mechanisms to deploy, to hire top data scientists, top computer scientists, top technologists, into the federal government.
You previously mentioned that a good analytics solution is 80% about finding the right data. With petabytes of data being generated, be it from satellite or the internet, what kind of analytics is required to sit through usable and relevant data and how can AI help in the process?
More and more data is being created every second, but more complex data is being brought to the table. What we found is that if you have 10 petabytes of data and you do analysis on it, you’re going to get some insight into what’s going on.
Talking about environmental justice, doubling the amount of data doesn’t provide double the amount of insight. At that point you could triple, quadruple or even 10x the amount of data. Still, you’re not going to get that much more insight from it because the analytics itself is limited in what it can find.
But through the power of modern AI, as you double the amount of data and amount of accelerated compute power, the amount of insights and the capabilities that we’re able to glean from this data grows linearly.
The more data we have access to by using AI to analyze, understand, and train on and then make inference on that data, we’ll be able to find linear levels of insight into it.
In the era of Geofencing, where everything is being driven by personalized data, how can we apply principles to it that are both robust and ethical, especially now with generative AI in the picture?
The short answer to that is that we’ve historically been in this space of exposing our data to large organizations that are managing our personal data; whether that’s photos that we put online or content like YouTube videos or blog posts. Even research work organizations themselves function as data repositories that store our data and then use it however they want in this new world.
In order to be ethical, fair, and manage that data correctly, we need to start understanding how we as individuals can manage and have oversight over our own data and share with entities as we see fit, as opposed to just giving it to other entities to do what they want with it. So I think there has to be a more granular.
Geofencing in terms of identifying specific not only areas but even granular spaces. We have to be even more granular in terms of how we’re storing and managing data down to the person.
From your experience at New York City’s data analytics department, what were the benefits of solving problems with data, and what can be the learnings for cities in third-world countries?
What I believe is also the case where it’s applied appropriately; where data, analytics, and AI are applied in a systematic manner. Through this, we were able to partner with city agencies to help them think through four key things.
How can they do their services faster? How can they do this work with scale? Because New York City is a big city and it’s really tough to think through.
It’s easy to think through how you do something, implement some sort of service on a block or in the neighborhood, but how do you do that citywide? So speed, scale, and then we had to work with them to be more accurate and precise in their work.
This is important in terms of providing services for residents. So we work with them and use data to help them meet the four criteria; speed, scale, accuracy, and precision. I think data and analysis and AI provide the key role for these criteria.
You’ve been very vocal about open and accessible data. Can you please talk more about the benefits of open data and the role of technology in it?
Similar to open source, open data does a couple of things. One, it provides a baseline of understanding. It’s the same thing we tend to say in social settings that any argument, any debate has to always start with the same set of facts.
You can’t have two sets of facts and then expect to get to some solution with regard to a debate. So what open data essentially says is as a core baseline for solving any problem, we have to all be working from the same understanding. Open data allows for that shared understanding.
But then also, the data and analytics side, to make sure that we have people who are from these communities that are adversely affected by climate and nature change that haven’t been engaged historically, that solutions have to come from these communities and not only environmental solutions but also the tech and AI solutions have to come from these communities that are affected.
The students, researchers, and even schools that exist in these communities are stepping up to offer solutions. The only way you can do that is by having access to the data that everyone has access to.
The top scientists at NASA have access to data that now these high school researchers in this one community have access to, and that’s what open data does. It connects those pieces such that you can bring more people to the table to provide solutions.
Quoting the Bezos Fund, “solving climate change requires more than lowering emissions”. Being the director of AI and data strategies, how can data aid towards a holistic climate-positive action around the world?
It is a combination of everything that has been mentioned yet. Understanding the problem leads to clearer solutions for solving a problem globally. And we know that these actions have to be taken in partnership with many organizations.
One country and one organization are not going to solve this global issue. All countries and organizations are going to have to solve this, which is why having a clear understanding of what is happening and then partnering up on how to solve that is important. Thus, shared technology and understanding are key here.
What is the roadmap ahead for the Bezos Earth Fund?
We are planning an AI program at the Bezos Earth Fund, where we have just launched our $100 million grand challenge for the climate and AI solutions.
So we are looking to work with and engage with organizations globally who are thinking through how to use AI to solve climate in nature, and really we’re our key focus is on bringing together climate and nature organizations with AI organizations to solve these problems in this decisive decade.
