India's Poor Left Out by Government's Coronavirus Contact-Tracing App

·8 min read

GURUGRAM, India -- In May domestic worker Husne Aara Bibi desperately approached one of her employers in the northern Indian outsourcing megacity of Gurugram, seeking a smartphone to download a government-developed contact-tracing and mapping mobile app tailored for the novel coronavirus.

Bibi's husband needed to show proof through the app that he was at low risk of becoming infected by the virus to be able to enter the luxurious apartment complex where he had for years cleaned cars until the Indian government ordered a lockdown for its 1.3 billion people in March.

Prime Minister Narendra Modi has urged citizens to use the mobile app Aarogya Setu. The government's push in promoting the app went so far as to make its use mandatory; in one Indian state police even threatened prosecution for non-compliance.

The government's official messaging touts the app as a reliable bodyguard. Lending support to that message are Indian movie stars, one of whom is featured in an ad for the app and popularized the tag line "Setu Mera Bodyguard," Hindi for "Setu is my bodyguard."

But the government push to use the mobile app is also stirring questions about its value. The app is tailored for smartphone use, and currently only about 500 million people in the country have such devices. So far about 124 million people -- less than one-tenth of the country's population -- are registered as users of Aarogya Setu; leaving more than 90% of the population from the purview of the app.

The GPS and Bluetooth settings of a smartphone using Aarogya Setu have to be on at all times so that it can automatically update the location data and pick up signals through Bluetooth when app users come in contact with each other.

"The utility of a smartphone-based proximity-based contact-tracing app may be limited by the number of people who have smartphones, who download the app, and who have the app turned on," says Ron Rivest, an institute professor at MIT and the lead principal investigator of the Private Automated Contact Tracing ( PACT) project.

[MORE: India Struggles to Maintain Social Distancing]

Privacy Worries Persist

The bulk of the country's registered users of Aarogya Setu access it on Android phones. A feature phone version of the app was recently launched on certain handsets marketed by Jio, a telecommunications company owned by Indian conglomerate Reliance. Personal details such as name, age, gender, phone number, and profession are sought to register with the app.

The self-assessment feature built into it asks users for their recent international travel history, symptoms of cough, fever, or breathing difficulty, as well as history of diabetes, hypertension and heart or lung disease. An interaction with someone who may have tested positive for COVID-19 also needs to be declared.

The government's privacy policy for the app promises users data security. However, the information collection process has left many uncomfortable while India awaits a data protection law.

"The basic worry with the use of the app is privacy and surveillance," says Apar Gupta, executive director of the Internet Freedom Foundation, a non-profit, that led the efforts toward making the use of the app voluntary. "There is policing of people through Aarogya Setu as they enter and exit office buildings and residential complexes."

A lawyer by training, Gupta says that despite his organization's efforts there is still confusion among citizens whether using the app is mandatory, which it is not legally.

[MORE: Top 10 Countries for Technological Expertise, Ranked by Perception]

A False Sense of Security

Eventually Bibi managed to get a free second-hand smartphone from the employer -- a new one is beyond the couple's means. Aarogya Setu is available in English as well as 11 Indian regional languages, including Bengali, Bibi's mother tongue. But that still doesn't help her understand how the app works. "Setu showed that someone 1.5 kilometers (0.9 miles) away from my house has corona(virus)," she says.

Bibi's sense of security is dangerous. She is unaware that the app only processes data it collects from registered users. That excludes most occupants in the three-story tenement she lives in because they cannot afford smartphones.

The risk of infection runs high for her and a vast majority of Indians who live in cramped settings where social-distancing is a luxury, toilets are shared, and masks and sanitizers are an additional expense. The loss of livelihoods in cities due to the lockdown has forced many migrant workers to return home. Some of them are taking back the virus with them to rural India which has a weak medical infrastructure and limited testing capabilities.

Subhashis Banerjee, a professor of computer science and engineering at the Indian Institute of Technology, says the app is not backed by a clear theory of electronic contact-tracing and lacks accuracy in predicting the risk of infection.

"What if you briefly come in contact with an infected person? How accurately does the Bluetooth pick up data?" he asks. "When the app shows that an infected person is, say, within a 500-meter radius, it does not provide a sense of direction."

On May 26, India's government made a further investment into the app by announcing a bug bounty and making public the source code of its Android version. This step had been long demanded by lawyers, tech experts and academics like Banerjee. Among those who welcomed the move was Baptiste Robert, a France-based security researcher, who uses the pseudonym Elliot Alderson on social media. He had engaged with the government on Twitter about a security and privacy issue with an earlier version of the app.

[MORE: Coronavirus Pandemic Shines a Light on the Curse of Caste]

Flaws in the App

Little attention is paid to India's low digital literacy level being a hindrance to the app's manual data collection process. Dr. Sushila Kataria, head of the Department of Internal Medicine at Medanta hospital in Gurugram, says that people have come to her anxiously after the home screens of their apps had turned Orange, which signifies a high risk of infection.

"On the basis of the color on their Aarogya Setu, I cannot test everyone," Kataria says. "Medical contact-tracing requires detailed questioning." She says, for example, that she does not know if colors on the app had changed because the users had self-assessed incorrectly. But she says that "the app is not proving to be very useful as India is nearing community transmission stage."

Her experience using the app has been anything but satisfactory. "Suddenly one day the app's home screen turned red for me," she says, referring to the color that shows the user is COVID-19-positive, an impossibility because she had not been tested around that time.

Despite wearing protective gear while treating her COVID-19 patients -- about 80 so far -- her risk of infection is always high. "I ended up wasting a testing kit because of the false positive on Aarogya Setu."

Lalitesh Katragadda, the head of product for the Aarogya Setu app, says such glitches should not occur as rigorous testing was done before releasing the app. "Results generated by the app are impacting the lives of people," he says. The potential of inaccuracies in the self-assessment process are factored in while using that data for an individual's risk evaluation, as well as mapping of potential hot spots of infection.

"Self-assessment is dodgy," Katragadda says. "Only about 60% users give accurate information." He acknowledges that there is a strong communication challenge in trying to make people understand that on the app, "green doesn't mean you are safe and orange doesn't mean that you are unsafe." Efforts are being made to improve this messaging, he says.

Contact Tracing Needs Health Infrastructure Support

Compared to the United States and Europe the number of coronavirus cases and the death toll in India may seem low, but those numbers are on the rise. As India is opening up and people are coming in contact with each other, the efficacy of Aarogya Setu will be tested.

"Currently, there is little evidence that DCT is effective to curb the COVID-19 pandemic beyond a theoretical assumption," says Dr. Soumya Swaminathan, the chief scientist for the World Health Organization, referring to digital contact tracing.

Swaminathan says that any contact tracing effort, whether digital or not, can only be effective if there is a strong public health infrastructure that includes adequate health services personnel, testing services and response mechanisms.

"Governments should introduce targeted solutions which consider the economic and social context of different groups," she says. "Efforts such as creating simplified, intuitive versions of digital interfaces for older generation, low-bandwidth versions for less connected parts of the population, or wearable solutions for people without smartphones, can all improve the penetration of a solution."

Pamposh Raina is a journalist based in New Delhi. You may follow her on Twitter here.