The Digital Brahmin explores how artificial intelligence is embedding caste biases into India's future. The analysis highlights the implications of AI systems that reflect social hierarchies, particularly through the lens of caste. It discusses the risks of institutionalizing bias in hiring, governance, and other sectors as AI technologies are rapidly adopted. This critical examination is essential for policymakers and technologists aiming to create equitable AI solutions. The work emphasizes the need for comprehensive audits to ensure fairness in AI applications, making it a vital resource for those interested in technology and social justice.

Key Points

  • Examines the impact of AI on caste biases in India
  • Discusses the risks of institutionalizing bias through automation
  • Highlights the need for audits in AI deployment
  • Analyzes how AI models reflect social hierarchies
  • Explores implications for hiring and governance in India
Advika Agrawal
6 pages
Language:English
Type:Article
Arts and Humanities
Advika Agrawal
6 pages
Language:English
Type:Article
Arts and Humanities
398
/ 6
EXPRESS OPINION
BECAUSE THE TRUTH INVOLVES US ALL
AI has learned caste — and caste
bias. And India is deploying it
faster than it can address the
problem
Whether AI 'knows' caste is not the most pressing question. Before these systems
subtly transform into infrastructure, determining who is perceived as employable,
credible, risky, or deserving, the question is whether India will measure what it
deploys
CELEBRATION
PARTNER
EDITORIALS COLUMNS
CO PRESENTED
BY
JOURNALISM OF COURAGE
The nation runs the risk of institutionalising bias as "automation" if AI is being acquired and implemented more
quickly than these audits become commonplace
Amidst the ongoing “India AI Impact Summit 2026”, a report highlighted a concern
that many researchers have been raising. If you give an AI model just an Indian
surname, it often guesses a person’s social background — sometimes linking
dominant-caste names to high-status jobs and Dalit names to less respected work.
India is not the only country with this problem. Even language models trained on
ordinary web text can pick up human-like biases, as demonstrated by Aylin Caliskan
and her colleagues in 2017. These are the same associations psychologists find in
implicit-bias tests. The argument was not that machines are biased, but rather that
social hierarchies are reflected in language, and if that is the intention of the models,
they will pick up on these patterns.
Written by:
Dhiraj Singha
6 min read
Feb 19, 2026 01:12 PM IST
First published on: Feb 18, 2026 at 11:25 AM IST
Real-world systems have shown this pattern. After learning to penalise signs
associated with women, like the word “women’s,” in club leadership roles, Amazon
discontinued an internal AI recruiting tool, according to a 2018 Reuters report.
Although engineers attempted to correct it, the key takeaway was obvious: The
system interprets biased data as significant signals if it is used for training.
Because caste is heavily encoded in text, particularly in surnames, it is ideally suited
to become a “signal” in India. Correlations such as which surnames co-occur with
“IIT/IIM,” “manager,” and “English-medium,” and which surnames more frequently
appear near “contract work,” “relief,” “scavenging,” or descriptions of deprivation,
will unavoidably be picked up by a model trained to minimise next-word prediction
error. It just needs enough frequent occurrences to understand what caste is.
These patterns are transformed into shapes in the model’s data behind the scenes.
Tolga Bolukbasi and his colleagues demonstrated approximately a decade ago that
bias manifests in the way names and words are grouped together, which can
reinforce stereotypes. They used the comparison between “programmer” and
“homemaker” as an example to illustrate how meaning can be arranged unevenly.
Since the math only follows recurring patterns rather than social categories, caste
can be represented in the same way as gender.
For this reason, research on India is crucial. A number of models covering social,
economic, educational, and political aspects are examined in the 2025 DECASTE by
Prashanth Vijayaraghavan and colleagues. It discovers that the addition of caste
cues, such as Indian surnames and personas, consistently reinforces stereotypes.
Another point that is frequently overlooked in public discussions is that safety
/ 6
End of Document
398

You May Also Like

FAQs

How does AI reinforce caste biases in India?
AI systems in India often learn from biased data, which can perpetuate existing social hierarchies. For instance, models may associate certain surnames with higher status jobs while linking others to lower-status work. This reflects broader societal biases encoded in language and data, leading to discriminatory outcomes in employment and other areas. As AI technologies are deployed without adequate oversight, the risk of institutionalizing these biases increases, making it crucial to address these issues proactively.
What are the implications of caste bias in AI for governance?
Caste bias in AI can significantly impact governance by influencing decision-making processes in hiring and resource allocation. When AI systems are used to screen candidates or assess creditworthiness, biased algorithms can lead to unfair treatment of individuals from marginalized communities. This can exacerbate existing inequalities and hinder social mobility. Therefore, understanding and mitigating these biases is essential for creating fair and just governance structures in India.
What measures can be taken to address caste bias in AI?
To combat caste bias in AI, comprehensive audits and assessments of AI systems are necessary. These audits should focus on identifying and mitigating biases in training data and algorithms. Additionally, developing context-specific benchmarks that reflect India's social structure can help ensure that AI applications are equitable. Engaging with diverse stakeholders, including marginalized communities, in the design and implementation of AI systems is also crucial for fostering inclusivity.
What role do surnames play in AI bias related to caste?
Surnames in India often carry significant social implications, including caste associations. AI models trained on data that includes these surnames can inadvertently learn to associate certain names with specific job types or social statuses. This can lead to biased outcomes in hiring and other areas where AI is applied. Understanding the implications of these associations is vital for addressing the underlying biases in AI systems.
Why is it important to measure AI's impact on caste?
Measuring AI's impact on caste is crucial to ensure that technological advancements do not reinforce existing social inequalities. As AI becomes more integrated into various sectors, understanding its effects on marginalized communities helps identify potential harms and areas for intervention. This measurement can guide policymakers and technologists in creating more equitable AI solutions that promote social justice.

Related