Digital safety of native Odia speakers: A socio-economic overview 

When compared to its dominant-language counterparts, the Odia language from eastern India had to wait for decades to be widely used on digital media. Today, Odia-speaking users of digital tools and the internet experience a series of digital safety and security risks. How are their experiences intersected by gender, race, and their geographical locations? How does limited access to information in Odia impact their digital security?

In a world where languages are constantly hierarchized, Odia language is both privileged and marginalized. On one hand, this language (and its speakers) dominate the neighboring languages and the native speakers in the region of Odisha, India. On the other hand, Odia speakers’ face very similar challenges to other low and medium-resource languages speakers, which are now experienced in the digital world, such as lack of representation or presence in media and the digital environment.

I am a native Odia speaker and I self-identify as a person with caste, gender and socioeconomic privileges. I also have professional level fluency with digital tools and platforms using Odia since 2006, an era when online communications of text typed in Odia was rare. My privileges also let me become one of the earliest and most active contributors to many platform interface internationalization localisation ^[3] and other research and development initiatives. I acknowledge and apologize for the role of my ancestors as oppressors, and my privileges that are an outcome of the historic oppression of the Dalit-Adivasi-Bahujan peoples. Though I put conscious effort in my work towards the abolition of brahminical caste-based oppression, I also stand corrected for any erroneous act of mine.

Based on my own position in the larger Odia-speaking community, I undertook the research  “Digital safety awareness of native Odia speakers: A socio-economic overview.” The study looked closely at various socioeconomic factors, such as linguistic and digital fluency, gender, caste, access to knowhow, and the ability to purchase devices and data, impact the users’ access to digital risks.

I conducted semi-structured interviews to listen to and better understand the perceptions and challenges from three individuals, each positioned uniquely in relation to their linguistic (in Odia) and digital fluency. First, Aliva Sahoo, who is a native speaker of the central (Mogulbandi) dialect of Odia. Second, Tularam who speaks Sambalpuri/Kosli, a language that is linguistically categorized under the Odia macrolanguage and, lastly, Prasanta who has multilingual fluency: his native language is Santali, he studied Odia and he speaks the Sundergarh variant of Sambalpuri. In addition to understanding their barriers, I also framed recommendations based on the findings that could help improve the current state of digital safety and security risks.

Online — and offline — linguistic privileges and vulnerabilities

Like most other places, digital platforms, applications, and services in India, make users who are already marginalized much more vulnerable to risks. The lack of vital information that leads to “information poverty ^[4]” for India’s poor is well documented.  Next I share my analysis and reflections from an intersectional perspective that considers caste, gender, geographical location.

Caste

The Hindu caste system is a 3000-year-old discriminatory culture ^[5] that assigns hereditary caste-identity based on ancient occupations. It divides people and imposes professions, each tied to varying privileges (or the lack thereof). Children also inherit the caste titles, often revealed in the surnames, from their parents, and inherit the dominance or marginalization tied to their caste. There are four main caste groups in order of social reverence: Brahmins (the priestly class people who are considered intellectual by birth), Kshatriyas (the warriors), Vaishyas (the businesspeople) and the Shudras (the “servitors”). Dalits ^[6], a heterogeneous ethnosocial group that includes peoples of different ethnicities, faith systems and from different geographical locations, fell outside these designated castes and were formerly considered as “untouchables.” Out of the 1.4 billion people in India, nearly 200 million are Dalits.

The British colonists took advantage of the pre-existing caste-based oppression to divide and rule, which helped the dominant castes, particularly the Brahmins, Kshatriyas and Vaishyas, to further carry out caste-based atrocities on Dalits after Indian independence in 1947. Tularam Bouddh, a Dalit himself, outlines how the historical socio-economic oppression has resulted in many Dalits owning smartphones much later than their dominant caste peers. The smartphone fluency of the Dalit youth is naturally delayed. Dalit activists self-censor and tone down their activism-related posts to invite less attention. They also experience a form of “digital oppression.”

Gender

Like the Adivasi-Dalit-Bahujan people, women in India have staggeringly limited access to information due to low linguistic literacy, limited access to mobile devices, and low knowhow of (digital) safety and security. This issue is much graver among low-income groups ^[7] and Dalit women, as nearly 100 million women suffer three times more oppression ^[8] due to poverty for being both Dalit and women. Adivasi women meet a similar or worse fate due to low literacy and access ^[9]. In addition to caste and income strata, patriarchy has grossly impacted women’s access to technology affecting their social and social media behavior. As I learned through the interviews, most elderly women from low- and mid-income groups keep themselves away from technology out of fear, which stems from a lack of knowhow, and the young women from such households often have a repressed digital presence.

Many women keep their digital presence and interactions to a minimum, fearing that more use would warrant more risks. When in trouble, many women also do not seek help from male companions for physical safety concerns. Aliva Sahoo, one of the interviewees, shares:

Honestly speaking, women might not feel very safe to ask their male peers for any help. It is even trickier with new acquaintances. I had to decline an in-person meeting offer when an acquaintance suggested one while speaking over a call.

Despite fluency in Odia, low English fluency restricts women’s digital movements as they often rely on other companions, mostly other women, to navigate the English-majority digital sphere. The lack of content, interface, and user documentation in Odia, makes women more vulnerable to digital risks while excluding them further from access.

English, Hindi and the linguistic domination

Regarding linguistic domination, Hindi is next to English in India as most platforms, starting with the public digital infrastructure and most private fintech platforms, prioritize interface localisation in Hindi. It is taught in schools in many Indian states where most people speak other languages. The domination of Hindi in popular culture, including through Bollywood films and television entertainment and news, result in speakers of other languages from many states, particularly Odisha, accessing content or using interfaces in Hindi. On the other hand, users from states with historical sociolinguistic movements for establishing state languages, such as Tamil, opt for English instead of Hindi as an alternative.

Geographical location and representation

Noted Dalit academic Raj Kumar has raised across his writings the dominant caste male writers from eastern/coastal Odisha occupying the literary space ^[10] and influencing the dominant caste-favoring shift of public narrative. He has also argued how creative writing by women, particularly from western Odisha, has been suppressed ^[11] by this literary community. A similar impact is also seen in Odia-language content in general. In formal writing, there is very little representation of different dialects and languages within the Odia macrolanguage, and, as a natural extension, the standard Odia digital content reflects this bias.

Tularam Bouddh, native Sambalpuri speaker (Bargarh variant), sees on one hand a rich vocabulary in the tongue he grew up speaking, but on the other hand, struggled to comprehend and write in standard Odia despite studying it in school. Prasanta Hembram grew up speaking Santali, and finds the quality of the Odia translation to be very limited across platforms and applications, especially the interface localisation and the terms and conditions. He tends to trust less such content, and switches to English. Recognising this barrier, he contributes to the localisation of interfaces of different digital platforms in Santali to make platforms more accessible for common users.

Learning from people’s experiences

The three interviewed users shared experiences related to digital vulnerabilities of their own and of their friends and families. Interface localisation, language fluency, digital fluency, caste, gender, economic strata are some of the important factors governing how these vulnerabilities affect people differently.

The larger dominant presence of Brahmins and other dominant castes in the content ecosystem further makes the digital world heavily influenced by Sanskrit, a language historically used by the Brahmins and other upper castes, that is incomprehensible to common people. This dominance has also slowed down the growth of the larger creation of literature, both digital and otherwise, related to the anti-caste movement and other social reforms. The dominance of the Central/Mugalbandi dialect of Odia in the overall macrolanguage content affects the former’s influence in the glossary for interface localisation. Digital interfaces and translated policies of platforms, software, and websites, wherever available, see an elitist register of Odia that is hard for people from oppressed caste groups and speakers outside coastal Odisha to understand.

The interfaces  currently available in the Odia macrolanguage are scarce and largely lack high-quality translation. This affects web users as they often come across misleading advertisements with clickbait titles in addition to spam calls and text messages asking for banking and other sensitive information. The vulnerability to spam and misleading content grows further for many people with lower fluency in English (as most ads and spam messages are still served in English).

Many who speak different registers of the Odia macrolanguage other than the Central dialect struggle to understand their interfaces. Though some platforms have consciously included localised Odia interfaces, high-quality and consistent standardization of localised terms and user guides are yet to be seen and this dissuades users from using such interfaces for long.

Users often form small groups for peer-to-peer knowledge exchange. The availability of videos on YouTube in topics concerning digital safety and security is becoming very helpful, even in languages like Hindi that are moderately understood by Odia speakers. These help users make more informed decisions about their digital interactions. Slightly more advanced users look information up online to cross-check it and they play the role of ambassadors to educate others in their networks.

Considering this complex scenario of access to digital safety, I suggest the following recommendations:

• Interface localisers need to be aware of the linguistic diversity and the resulting needs of users who are less fluent in the Central dialect of Odia, to create informed localisation.
  
• The localisation needs to take into account colloquial and other commonly-used terms such as technical or medical terms, including loanwords, as opposed to vocabulary influenced from Sanskrit or elitist neology. Input methods ^[12] need to be simpler and easier, offering more user documentation and on-interface support such as predictive text ^[13].
  
• Voice-based Human-Computer Interactions can be a better alternative to only text-based interactions while keeping in mind the ethics at large and, more specifically, consent for collecting and using voice data for machine learning.
  
• Audio content on digital platforms is much needed to bridge the literacy barrier. Information related to natural calamities, health and other emergencies should be made available in multiple formats.
  
• Robust misinformation reporting mechanisms need to be enforced on platforms. This also includes digital content with misquoted famous personalities. The reporting mechanism needs simplification to accommodate the needs of users with basic fluency.
  

To create a truly inclusive and equitable digital world, it is imperative that the deep-rooted socioeconomic and digital oppressions of several marginalized groups in Odisha are abolished. This would require affirmative action to dismantle the caste system, patriarchy, and economic inequality. While there are recommendations for immediate technological reforms to make digital experiences of marginalized groups more inclusive and accessible, lasting change will require a concerted effort from social justice movements, public discourse, and the creation and distribution of literature on these issues. Only by acknowledging and actively working to dismantle all the oppressive systems while building ways to increase the active participation of historically-oppressed groups can help improve the state of digital safety and security.