We Count, Counting Ourselves

Many of the world’s poorest citizens are invisible in development data. During the COVID 2nd wave in India, journalists and policy-makers scrambled to understand how bad COVID is in rural areas. 

We Count, Counting Ourselves is an AI powered layer of Video Volunteers hyperlocal network of citizen journalists in India. The data collected will be made available into data registries and we will integrate the key facts into hyperlocal social media conversations.  This will enable solution-taking, provide timely data to experts and bridge the communications gaps that hamper citizen-state synergy. As of 17 July 2020, UN Stats has designated 106 out of 129 SDG indicators as Tier 2, meaning '...data are not regularly produced by countries.' Much of the data towards those indicators can be gathered by citizens. When our solution scales globally, it will enable citizens to hold governments to account for maintaining progress on the SDGs. 

There is not enough COVID (and other) data collected in a systematic way from India’s rural areas. Anecdotal community-led stories have emerged and directories created, however the system lacks the layer of trustworthy community hubs that can activate and collect the data in a way that can be effectively used by the government and international agencies, and referenced by journalists and activists.

This has accelerated COVID’s spread, by preventing aid agencies from allocating scarce resources like oxygen to the most needy blocks/districts. 

Much missing public health data could be provided bottom-up, by citizens themselves, e.g, the quality of the village public health center; # of people with COVID symptoms; # ambulances; hospital wait times. But citizen-generated is not currently used by any govt agency. 

A national citizen-generated data ecosystem doesn’t exist because of the historical indifference of authorities in listening to/valuing the voices of the poor, prejudice against the trustworthiness of citizen-generated data, and, until recently, poor digital connectivity. 

It’s also been difficult to access rural citizens at scale to enlist them as enumerators, or to penetrate their WhatsApp groups. The rural poor are mostly on WhatsApp, not twitter, so NLP to gather citizen’s data is almost impossible. 

Our solution is to scale the impact on government functioning of our existing hyperlocal network of grassroots citizen journalists across 200 districts through AI chatbots. 

They will recruit data collectors near them to gather data on COVID. This will happen via an-easy-to use surveying chatbotWhatsApp (widely used by 400m+ people in India and increasing) that will push out many COVID data questions each week to gather millions of data points. This is based on our 15 months’ experience providing/influencing COVID relief across 200 districts and making these communities ‘COVID prepared'.

Through grassroots NGO partnerships and social media virality - responding to citizens’ healthy desire to monitor government -our application will become popular across the country, as citizens take our surveys. An additional bot present in the whatsapp groups will support our Correspondent to identify COVID misinformation in her whatsapp groups so she can respond. It will also monitor and classify utterances to gather further citizen-generated COVID data.

Our solution improves the ability of government, civil society and market actors to integrate and solve together & manifests two characteristics - Emitting (ie broadcasting insights for immediate visibility via social media & journalists who take notice & amplify) & Building (for long term sustainability). 

We are working to directly and meaningfully improve the lives of rural and peri-urban households in India that are currently not receiving government programs they are intended to receive -including health services like COVID testing or adequately staffed local health clinics. This population - of those who are currently invisible to the particular programs meant to serve them- is huge. $122 billion is allocated for over 1000 government anti-poverty schemes each year, but much is unspent each year, in part because it is extremely difficult for citizens to escalate a grievance, ie, make themselves heard, when they are unfairly refused access to a government program. For instance, a survey found that 40% of the government’s Jah Dhan accounts could not access the government’s COVID relief. This program is directed at approximately 200 million women, which may mean that up to 80 million are unable to access it. This is what it means to be voiceless vis a vis the government. This solution will meaningfully address their needs by enabling their own village-level data on health crises and health infrastructure (as well as other poverty-related issues) to quickly, regularly and in an organized fashion to come forward. This data will relate to 'data gaps,' ie, areas where the data is either absent or not usable/accessible by citizens. Armed with such data, the communities and their advocates/allies in the government, NGO world, media and CSR space will be able to work together and co-create solutions and thus improve citizen-state synergy. 
 
As India's largest community media organization, Video Volunteers deeply understands their needs - we have facilitated over 8000 community videos to date, all of which deal with community problems & solutions, based on the principle that people have a right to speak for themselves rather than be spoken for, in their own articulate, and unmediated voice. Filling in data gaps is core to our work - since 2006, we have monitored the implementation of the RTE (education) act and maternal health services, provided granular data across many districts related to electrification and ‘ground truthed’ the government's statement that it had made India 'open defecation free'. In our program 'Surveys for Action' we have gathered in depth quantitative data on numerous issues, including COVID preparedness and the spread of vaccine disinformation, at a time when the news media was paying no attention to these two issues. Within 200 WhatsApp Action Groups that we have set up, we are listening to and engaging with communities around COVID solutions. With the new partners on board for this project and the technology solution (the Chatbot), we can scale our model of enabling communities to fill data gaps to the entire country. 
 
The scalable tech solution described here, which has been developed as part of the ASPiRe Program led by Societal Platform and Ashoka, will be anchored by our 200 Community Correspondents, who come from India's most marginalized communities (30% Dalit, 45% adivasi, 50% women). A selection of them, who are part of the VV Council, will be deeply involved in designing the ChatBot/platform roll out to make it effective. 

Our community-generated data system will be able to slow future disease outbreaks by facilitating last mile data collection from the remote villages about which it is very difficult to gather data. This and other articles make clear there is agreement that we don't know the extent of the crisis in the rural areas due to poor data. 
 
In addition to the surveying chatbot, a conversational chatbot created with guidance from AI scientist Dr. Rohini Srihari will be used to combat vaccine disinformation in 200 WhatsApp Action groups VV already runs, and thousands more accessed through NGO partners. 

VV is a sustainable enterprise working in 200 districts of India, focused on community-generated data in multiple forms - video (8000+ videos on our YouTube channel); in-depth quantitative surveys (several done each year since 2015), 200 village-level WhatsApp Action groups, plus qualitative data on thousands of our local action campaigns (called ‘impact’ processes) which we have analyzed with our research partner, as well as hundreds of social media campaigns. Though the Chatbot and the data commons outlined here are new technologies for us, we have selected ‘scale’ because they are extending our existing work and will help us be more efficient in enabling communities to monitor government programs and organize for solutions. In the coming years, all of those efforts will be unified on our data platform, and the Chatbot -- perhaps the most scalable, efficient community media tool imaginable --will enable millions of citizens to provide actionable, organized data. 

There are data registries and spreadsheets and twitter hashtags that helped citizens organize info on bed availability in April/May, but these are not maintained & not built for long term use. There is no effort to use surveys systematically to gather data direct from citizens on infection, bed/oxygen supply availability, number of deaths.There are efforts to gather citizen data in the healthcare industry (apps, monitoring devices), but these are for health care products. News articles used only anecdotes to report on COVID deaths, or proxy indicators like funeral fires.

The only available data from the ground on COVID is through two Government agencies in India (the Ministry of health and family welfare, and Indian Council of Medical Research.) Various reports suggest that this data is inadequate and lacking and not timely, which highlights that resources are not being directed efficiently. Because there is no citizen data being collected systematically, not one of the top sources cited for COVID data by the WHO is a citizen-created database.

Listening to and learning from communities, for the purposes of their own self-empowerment and to strengthen bottom-up democracy and decision-making, has a long history we are building on- ref. Paolo Freire in the the 60s, participatory rural appraisal and community mapping. So why isn’t it more common now? Because it’s difficult work. Our innovation - using a chatbot embedded in WhatsApp - will make citizen-generated data much more scalable. 

In the previous fiscal year, VV positively impacted 8.1 million people, meaning these are the number of people who directly benefited when we managed to resolve an issue that a community member solved from the data we gathered. 

In one year, it will be 10 million.
In five years, it will be 50 million.

(Note - these are numbers for all programs VV runs to empower people with a voice including this data platform.)

Current impact measurements are number of videos created; our social media reach; number of people affected by the issues raised in our videos; and most importantly, number of lives affected by the ‘impacts’ our Correspondents are able to make when the local campaigns they instigate, as they produce their videos, solve the problems the videos are about (8.5 million last year). We have devised an impact measurement framework with academics from UVA and the MIT Gov/Lab, that involves watching each video and speaking with the Correspondent to gather more than 100 data points regarding impact. This data is then used by these research partners for their larger research agendas around what makes citizens take action. Our database was described by YouTube Social Impact at one point recently as the largest video archive on the internet documenting village-level social change and social justice. Since that time, we invest tremendous time into analyzing what communities are talking about and presenting that to the world. 

There are 28 staff and 215 Community Correspondents (contractors) in the VV Network

Jessica, Tony, Dani, Susan, Ashoka & Societal Platforms. 

Jessica brings a rigor around accurate information that comes from her news & media background, and that will be important in this project. She has lived for nearly 20 years in India working with communities to empower them with a voice. Ashoka and Societal Platform, with whose mentorship this proposal was developed, are world experts in what makes a social solution scale and how to build digital platforms. Our two advisors and likely partners for the implementation of the AI/Chatbot solution, Donald Lobo of Glific and Rohini Srihari of the University of Buffalo are technology experts with deep experience in the social sector in India. VV’s strength is that it is deeply rooted in the grassroots, with an on-the-ground presence in 200 districts, plus deep international experience in networks related to technology, social change and the corporate world. Several of VV’s staff who will work on this project themselves come from the grassroots, as VV has a policy to hire for positions where possible from within the ranks of Community Correspondents.

Starting in 2018, VV created a unique leadership body called the VV Council. It is made up of currently 25 Community Correspondents and 7 staff. Currently functioning as an Advisory Council, in time, the VV Council will hopefully become a legally registered cooperative, a first of its kind in grassroots journalism. The Council currently works through five committees --safety & security, communications, impact, Correspondent Welfare, Activation --as well as a Core Committee. The organization has acted on the advice of the Council in numerous instances, such as changing payment structures, insurance reimbursements and smoothening processes around distributing equipment. At its last in-person meeting prior to Covid, the Council charged the organization with working towards the goal of ‘having the heart of a social movement, combined with the professionalism and accountability of a business.’ The Council has made VV far more resilient to COVID, and has brought around programmatic impact such as higher rates of retention of Correspondents whose grievances are addressed by the Council. 

For us, SOLVE is an opportunity that provides access to mentorship and guidance from public health and statistics experts, a wide community of peers and technologists. It also opens us more doors for funders and other partners who can help us co-create an effective, sustainable and viable product/platform to meet the growing data gaps in India and empower a wide range of actors by giving them access to this Citizen generated data.

Our efforts at the grassroot community level across 200 districts in India for nearly 2 decades, and experimentation with gathering better data has allowed us to learn invaluable insights. Our experimentation and multiple discussions with various stakeholders and experts have helped us identify that chatbots could play a catalytic role in transforming our work. We seek your partnership and support with developing the technology vertical of our solve to scale up our efforts and to transform our insights into action. We wish to co-design and generate better monitoring and evaluation frameworks with your support and finally need your assistance and extended networks’ guidance for the business model vertical of our solution. This is the last piece of the puzzle as we aim to make this scalable, sustainable and importantly help our community workers with better livelihoods. Exploring a better business model with your support, we are confident will lead to better impact on ground. 

We seek to Partner with the Statistics and Data Science Center and the Public Health faculty at MIT for resources and support. We also wish to partner with the solver Mapsights (Aditya Sharma and Team of Sparrow Analytics), the spatial and temporal predictive analytics Decision Support System (DSS) for public health and governance - running in India. Apart from these, we wish to partner with Johns Hopkins Coronavirus resource centre, Ministry of Health and Family Welfare of Govt. of India and Indian Council for Medical research. We are already partnered with the MIT GovLab and would seek to involve them at some stage in this project.

Our citizen data gathering efforts will significantly rely on data science for the first year of our solution to benefit communities and create meaningful impact for them. All our data collected will undergo extensive data systems and algorithms to extract knowledge and insight into problems of the community to help us work towards solving it and/or build capacity for other stakeholders such as think tanks, nonprofits and CSRs to use our data to help the communities and rural people. Over the next year, we will aim to incorporate AI and ML to reach better scale and for in-depth data gathering and impact creation.