A: Checkpoint is a research project at PROTO, commissioned and technically assisted by WhatsApp. It uses a tipline on WhatsApp (+91- 9643-000-888) only as a means to collect information that is otherwise inaccessible given the nature of private messaging. WhatsApp has also contracted Meedan to expand the Check platform and integrate it with the WhatsApp Business API for use in this project, including to help manage the workflow from receipt of a query to sending back its verification report. Dig Deeper Media was contracted to provide consulting on frameworks design for the initiative. Learnings from the project will be shared with the International Center for Journalists for distribution across their global network.

The goal of this project is to study the misinformation phenomenon at scale — natively in WhatsApp — during the Indian elections. As more data flows in, we will be able to identify the most susceptible or affected issues, locations, languages, regions, and more. The verification reports we send back will encourage users to participate in this research as “listening posts” and send more signals for analysis. The final output at the end of four months will help WhatsApp determine next steps against misinformation on the platform.

A: The Checkpoint tipline is primarily used to gather data for research, and is not a helpline that will be able to provide a response to every user. The information provided by users helps us understand potential misinformation in a particular message, and when possible, we will send back a message to users. We would like to verify every rumor but we know that will not be possible given the diversity of information we will receive and the limitations of any verification research.

When we provide a response to a user, we are providing a verification about the reliability of the original claim. This is not as intensive as the fact-checking that a journalism organization can provide. Here’s an excerpt from the Verification Handbook that might help with this distinction:

Fact-checking is a specific application of verification in the world of journalism. Verification is a scientific-like approach to getting the facts … practiced by many people who are not fact-checkers — or journalists, for that matter.

A: There are enough algorithm-enabled projects studying misinformation on open networks. Checkpoint's research is using a WhatsApp tipline to crowdsource data that would otherwise not be accessible. Over the next four months, we expect to aggregate these signals at scale, to better understand how misinformation during large events of public interest in India — such as the elections — spreads across languages, regions, even issues. Our dataset, findings and analyses will be compiled into a first of its kind report for India that will be shared with WhatsApp.

A: A small team of analysts with backgrounds in journalism, research, technology work a 40-hour week at PROTO to meet the goals of the project — crowdsource misinformation, triage queries to identify those that are out of scope, verify those within scope and convert them into a structured database for retrospective research and analysis. We recognize we will receive far more submissions than we can ultimately verify and respond.

A: This project is the first of its kind in India and, for now, our goal is to understand the spread of misinformation during the election. We have not made any decisions about what happens after.

A: Yes, it is free. Users are not charged for forwarding a suspicious WhatsApp message to the tipline.

A: Users can send a rumor — including a text, image, or video link — to the tipline. They will receive a reply that the verification center will be reviewing the rumor for accuracy. If it is a known rumor the user will get a quick response. If it is a new rumor, it will take some time for the verification center to determine if it can be verified.

This project is made possible by a combination of emergent technologies that do not support direct video uploads right now.

A: The user will receive an image or text in response identifying the rumor as true or false. The message will include a link that will provide more context on the rumor itself.

A: The response time with vary based on the complexity of the submissions however verifications will not be instant. Users can send a rumor — including text, pictures or video links — to the tipline. They will receive a reply that the verification center will be reviewing the rumor for accuracy. If it is a known rumor, the user will get a quick response.

If it is a new rumour, and is determined to either be beyond the scope of this research or not verifiable based on publicly accessible information, users will get an “out of scope” response.

If the new rumour is both within scope and verifiable, the verification center will prioritize requests based on their urgency. Finally, if the new rumour is both within scope, verifiable and prioritized, the verification center may take up to 24 hours to send back a report.

The longer we run the project, the more robust our database and response times will be. Of course we couldn’t have a database without starting the project, so this is an effort that will take some time to calibrate.

After 24 hours, we will prioritize newer queries. It is not possible to respond to all queries.

A: The verification center has instituted a Standard Operating Procedure to make these determinations.

A query may be considered out of scope for a range of reasons. For example, if it’s satire or humour or just not related to elections, or if it’s in a language other than the five assigned to this project (English, Hindi, Telugu, Bengali, Malayalam), or if it cannot be evaluated merely on publicly available and accessible information, and would require journalistic reporting or fact-checking.

A: If an incoming query has already been verified before, then the software auto answers it. Apart from this, all verification tasks are manual.

A: WhatsApp is a private messaging platform and we cannot see the content shared between users.

A: This project is using the WhatsApp Business API, which enables organizations to receive user messages at scale. We have structured this project in such a fashion that both PROTO and Meedan do not automatically receive the phone numbers of users submitting information. These teams will receive rumors and may share the misinformation database with other organizations contributing to the project. Meedan works with a solutions provider called Smooch to provide its access to the WhatsApp Business API.

A: When a user sends a rumor to the tipline the content will go into a database maintained by Meedan.