From Pessimism to Promise, page 11
The reliability of these tools also came into question. Women sweepers wearing bangles complained that their watch would turn on when their bangle accidently touched the watch. Some complained that the watches showed the wrong location, causing a loss of wages. Their salaries could decrease if the wearable’s batteries died while they were at work, failing to capture their time.
Few redressal mechanisms and high demands on the workers’ time to address glitches result in more precarity and exploitation. As the private sector experiments with new tracking and automation of their workers, the public sector experiments with new ways to deliver essential safety nets, albeit with mixed results.
Online Safety Nets
In 2021, the Indian government launched the e-Shram portal, a national database of unorganized workers.28 The portal registers unorganized workers and connects them with social security benefits if they meet with an accident, become disabled or unemployed, or die at the work site. This is an unprecedented effort to formalize the massive informal economy.
Sociologist Shweta Mahendra Chandrashekhar worked for FemLab with a focus on the construction business in India. She interviewed union workers and leaders, municipal officers, and contractors to understand worker benefits, including how digital registrations improved access to welfare benefits. Her father owns a tunnel construction company in Pune, and her childhood was shaped by her visits to the construction sites and her interactions with the site engineers, migrant workers, contractors, and government officials. The COVID-19 lockdown hit the construction sector particularly hard since a majority of their workers are migrants. Pune saw about 75 percent of workers leave for their villages when the lockdown was announced overnight in March 2020.29
A majority of migrants failed to access welfare benefits, despite the government making funds available during this humanitarian crisis. Union leaders provided reasons for these failures: Migrants couldn’t directly register online, so contractors needed to register workers on their behalf. Sometimes the contractors themselves were unregistered and were subcontracted by other private contractors. Middlemen have few incentives to register migrants, as laws such as the Interstate Migrant Workers Act demand that they provide migrant workers and their families with health care, housing, education for children, semiannual train fare to their villages, and other benefits. The president of the Construction Workers Union, while hopeful about digitizing the registration system, argued that we need to simultaneously build awareness of these schemes and provide free digital literacy support to enable online registration. He remarked that otherwise, the digital tracking process could create another layer of middlemen and more opportunities for exploitation.30
The government’s efforts to standardize registration and simplify migrants’ access to benefits across states are fraught with challenges. These goals require consolidating migrants’ registered accounts, including their pension and other accrued benefits. The idea of a singular professional identity doesn’t match the reality of migrants who are seasonal farmers and work simultaneously in construction, ride-hailing, and domestic services to make ends meet.
Women face similar issues of being absorbed into the state welfare system. Their situation is compounded by the patriarchal culture in which being too public, too outspoken, and too present can have adverse consequences. Monitoring apps can amplify the social pressures women face and further deter their participation in the workforce. According to the World Bank, the percentage of women working in India decreased from 2010 to 2020, dropping from 26 to 19 percent.31 In fact, this decline has become a global trend, especially among female gig workers. Women are increasingly abandoning the use of digital tools and the internet itself as fear of a digital presence overtakes them.32 Locked profiles, while an option, keep them from capitalizing on the attention economy. Tracking devices deter movements. Social expectations relegate them to care work, which remains undervalued and underpaid. In such an ecosystem of withdrawal, what will it take for women at the bottom of the supply chain to go online and use networked tools to their advantage?
Not Just a Woman
Women behind the wheel.
FemLab’s media researcher, Pallavi Bansal, and I designed a project to gain insights on why it was so difficult to find female drivers for ride-hailing companies despite the high demand. We found double standards for women working with strangers in public spaces. Sushil Shroff, director of Taxshe, an exclusive all-women driver-on-demand cab service running in Bengaluru, explains that cultural norms result in female drivers being judged harshly and can deter them from choosing this as a profession. Shroff explains that there is a tendency to judge women drivers as being “cheap and available,” because “anyone can sit in her car and talk anything to her, and she’ll have to take it.”33
Taxshe had to rebrand women drivers as “alternate moms” to attract women to this profession, get their families to agree to them working in the ride-hailing sector, and shift customers’ mindset. The company leveraged the cliché trope of caring women that feminists typically want to break out of. They labeled female drivers as “Roos”—referring to kangaroos—who will protect young ones while moving. Shroff explains that they had to try to convince clients to respect drivers by equating them to mothers: “We told each of our clients, ‘like you, she’s a mother. Yeah, she’s handling her kids, and she’s driving your kids. . . . It’s a parent who’s driving. It’s not just a woman.’ . . . So that gets the respect.”34
Nonprofits like the Azad Foundation and the Neeva Foundation explain that it is essential to invest in female drivers to prevent them from dropping out. Women may have never driven a car or even opened a car door for a customer. They may not be used to talking with strangers. Filling out forms, getting a license, dressing professionally, handling finances, and dealing with the mobile app takes time and, more importantly, confidence. One driver, a widow with two daughters to support, shares her story:
I was very afraid of starting new things . . . my confidence level was not that good. First, they give you the confidence over there. Like “You can do, you can do,” that one is the first thing. . . . I was so afraid of everything. . . . I hesitated to speak, I hesitated to approach people, like that. That is one thing I am not now.35
While these organizations build women’s confidence and capacity as drivers, some fixes are more related to urban planning, social norms, digital affordances, awareness, and media literacy.
Empowering “Partners”
Women drivers at Ola and Uber express frustration about the lack of access to clean and safe toilets. Ola Mobility Institute, a think tank focusing on mobility innovation in India, explains that it is also the lack of awareness of app options that contributes to this plight. It has conducted surveys among women drivers and discovered that few are aware of how to navigate their app to find the nearest restroom.36 Additionally, the app could benefit from a rating system on the cleanliness and safety of restrooms for women. The SOS feature on the app, while present, is rarely connected to a live operator and often results in no response. Women drivers want to be able to rate their customers, too, and have their company aggregate these ratings. They want the company to act on the data insights and block customers who have broken rules of engagement with persistent rudeness, harassment, and overall bad behavior.
Women workers or “partners” with home-services gig platforms like Urban Company have shared similar concerns and desires for the ratings to be a two-way process with our FemLab researcher Sai Amulya Komarraju.37 Moreover, while they recognize the risks of everyday digital surveillance, they also see in them an opportunity to prove their worth and distinguish themselves. They feel pride in completing the ten-day training course, given that some fail in the onboarding process. They learn how to sanitize their products, deliver their services, present themselves, greet customers, operate the app, and check their commissions and ratings. The women learn the vocabulary of this digitized sector—power bank, app, leads, credits, ratings, reviews, automated, tracking, app store. These terms become part of their standard operating procedures. The partners feel professional and respected, and the data proves their standing to the company and customers alike. A service provider confidently explains the rating system:
Now you [the customer] will be questioned, how was the pro who was sent to you, did she behave well, did she sanitize, wear PPE kit, gloves, mask. If you see me doing all of this then you will tick yes, but if I have never used any of these things, you have never seen me use these things you say no.38
Many feel pride in being able to manage expectations, comply with hygiene standards, and maintain a high average rating in the system. Monitoring and rating systems can be repurposed as tools for transparency, company negotiations, promotions, and social recognition. The professionalization can have a spillover effect on the standing of these women workers in their community:
The moment you say beautician, to be honest, the mindset is they look at you differently, not with good perspective or good intention. Of course, this does not mean that everybody is like that. But UC [Urban Company] is a brand, so in the starting, during the training itself all of this is clarified. If you are working in UC, then people should automatically differentiate between you and whatever preconceived notions they have about beauticians, they must perceive us as thorough professionals. There is clarity about this now for sure.39
Building surveillance of care requires negotiating deep-seated cultural systems and infusing power into given labels such as “partners” by appropriating tools at one’s disposal. However, this is done through soft power, subtly increasing agency from within, and starting with a core power unit—the joint family.
Family Planning of a Working Kind
Meesho, the reselling app, learned this cultural norm quickly as it scaled its business model. Engaging women resellers on their platform is a family effort. Achyutha Sharma, the company’s user research manager, explains how over time these compliances lead to subtle but important shifts in the women’s status, freedom, and confidence in and outside the home. Sharma points out that at the start a woman must negotiate “social permissions” with her husband or family decision makers to get a mobile device and internet access for work. The family typically agrees on the condition that she also fulfill her household duties. Over time, as they see profits coming in, things begin to change in the woman’s status in the house and her bargaining power, leading to a higher degree of “digital confidence.”40
Family surveillance can ease over time to a healthy inattention that allows women to become more visible and vocal online. These tactics, while commendable, underline the weight of the patriarchal apparatus women need to push against to have a fighting chance in an increasingly automated workforce. The future of work will continue to be upended by disruptive technologies. Workers at the bottom of the value chain bear the heaviest burden; this is especially true of women in the informal sector. Media debates position labor futures as a contest between tech and humanity. Tech, however, is human. The digital is physical, built on the continuing sweat of workers’ experience. Algorithms learn from people as much as people learn from them.
Building institutions, infrastructures, and initiatives of care such as clean and safe toilets for women, two-way rating systems, confidence training, upskilling, and universal basic incomes can sand down the hard edges of patriarchal norms. These changes start from a position of compassion. Installing the surveillance of care demands a shift from the rational to the emotional. Feeling can become fuel to action, unless it is commodified against those that need the most change. Machinery and empathy do not have to stand against each other in this pursuit of social change.
Emotion 3.0
Machine learning is built to identify and structure opinions, emotions, and beliefs from collected text, expressions, and other biometric cues through wearable devices. Businesses valorize emotional intelligence (EI) to help build relationships, manage crises, communicate effectively, and inspire their teams. It provides a competitive edge. A 2019 survey with managers in the United States showed that 71 percent of employers favor the so-called emotional quotient (EQ) over the intelligence quotient (IQ) in employees.41
Empathy and social skills trump logic and problem-solving as management gurus tout EI as the key to good leadership. Impression management is all the rage. Recruiters are increasingly giving weight to EI when hiring.42 Insurance plans, advertising campaigns, product designs, banking protocols, health care tools, education content, and agricultural innovations are increasingly tailored to their diverse customers by mining their sentiments. This outlook was far from the norm in the past: emotionality was perceived poorly in society and was associated with women and the Global South. To be emotional was to be weak.
Personal Polygraph Machine
Feminists have long battled the undermining of emotion and its link with the feminine to keep women away from political life. A persistent trope of women revealing their emotions easily, like a “personal polygraph machine,” has worked against them as leaders are expected to be self-contained and self-controlled.43 This degradation of emotion applies to nations as well. The international peace organization Service Civil International, in their “Picturing the Global South” toolkit, highlights how countries in the Global South are typically perceived as “emotional” by those in the North.44
In the colonial days, imperial settlers framed Southern cultures as “emotionally pathological.”45 They viewed the habits of their new subjects as the antithesis of the Enlightenment values of reason, scientific progress, and technological order—the makings of masculine sensibility. Their worldview provided a “rational” justification and moral responsibility for settlers to “civilize” the Natives. Anthropologist Ann Laura Stoler explains how colonialism was driven by an “emotional economy,” where rulers strove to predict and prescribe what sentiments to perpetuate and what emotional contagions to contain.46
Until the late nineteenth century, emotion connoted more a collective than an individual state of affairs. It alluded to groups driven by feeling to action; as gender studies scholar Jacqueline Holler explains, people were “powerless to resist.”47 Inspired by the church and politically vested interests, emotional standards were established to manage and control people’s sentiments through rituals, institutions, and cultural practices. Settlers validated certain types of emotion, steered by beliefs about patriotism, racial superiority, paternalism, and honor.
Today, as emotional intelligence meets artificial intelligence, how do we negotiate the potential of collective care and the fear of collective control? As Stoler argues, “Sentiment is the ground against which the figure of reason is measured and drawn.”48
Age of Sentiment
Cultural anthropologist William M. Reddy recognizes emotion as the driving force of the Renaissance, the “age of reason,” and has retitled the era the “age of sentiment.”49 He may have inadvertently captured our contemporary time. The battle for the future of our sentiments as an instrument of care versus control has begun.
Meta has been inundated by controversies over its experiments using sentiment analysis. Studies conducted as early as 2014 by its Core Data Science Team and Cornell University’s Departments of Communication and Information Science found that “emotional states can be transferred to others via emotional contagion, leading people to experience the same emotions without their awareness.”50 In June 2022, Microsoft announced its new policy to phase out public access to several of their “emotional recognition” tools (such as Azure) after criticism of their flawed results.51 Experts argue that without accounting for context, facial expressions cannot be equated to internal feelings.
Spotify appears to have “solved” some of the context issues with its 2021 patent on monitoring users’ speech to infer their tastes.52 The AI captures intonation, stress, rhythms, and other elements of speech to assess emotional states. Machine learning contextually situates these insights with identity-based metadata (such as gender, age, and other such variables) and environmental metadata of users’ locations and physical surroundings to get a more accurate analysis of the emotions driving music choices.
During the COVID-19 pandemic, Zoom became a living room for billions of people and a global social experiment on people’s emotional analysis. Zoom analyzes users’ facial expressions, vocal patterns, body language, and gestures to deduce collective human behavior. In May 2022, Fight for the Future and twenty-seven other human rights organizations wrote an open letter to Zoom, calling for the halt of their AI-driven emotion-tracking software.53
While the efficacy of sentiment analysis continues to evolve, the intent to weaponize collective emotions remains a serious concern. How can regulators break away from these binary choices to steer tech toward a care-based approach? Can tech companies reverse engineer these tools to benefit society?
Reverse Engineering for Care
Tech companies must be pushed into action. Meta can build on such analytics to reduce sexual harassment, foster mental health, and build community. Spotify can repurpose itself as a music therapy portal and provide comfort to millions with better targeted playlists. Microsoft’s Seeing AI, a tool that helps blind people record their experiences and events, can be enhanced with sentiment analysis to deepen users’ recorded memories.54 Zoom can continue to build on its living-room feeling by designing for contagious joy, shared grief, and solidarity, as we witness online weddings, funerals, birthdays, graduations, and other events that mark our lives.
