Humans in the Loop is becoming central to India’s approach to Artificial Intelligence (AI) as the country’s IndiaAI Mission marks a strategic shift in how AI is woven into economic growth, governance, and social inclusion. As a frontier technology, to make AI a companion in India’s growth story, AI models need a human workforce to train, validate, monitor, and ethically guide AI systems.
This is where the Humans-in-the-Loop (HITL) model mandates our attention to look deeper into AI accuracy, reducing bias, localising AI for Indian languages and contexts and ensuring responsible and trustworthy AI.
Two key pillars stand out in the India AI Mission: one, inclusion for social empowerment; and two, sustainable models of human capital development. With India’s commitment to make AI a companion for work and living, the need for skilled talent is growing rapidly.
Shaping the intelligence of machines and training the algorithms
AI jobs are not just for engineers and tech experts. A significant share of opportunities exists at the base of the employment pyramid, particularly in roles such as data annotation and related services. At the core of an AI revolution lies the need for large training datasets, often unique to businesses. This need is addressed through data labelling and data annotation, creating scalable entry-level digital jobs.
The data annotation market in India can exceed USD 7 billion by 2030 and has the potential to engage a workforce of up to a million through gig and full time jobs. (NASSCOM Report)
Data annotation involves identifying, labelling, tagging or annotating raw data such as text, images, video or audio to enable machine learning (ML) algorithms to interpret and learn. It supports AI systems through structured, high-quality data, teaching algorithms to identify patterns, recognize objects, and make accurate predictions, directly fueling model accuracy and reducing bias in AI output.
These AI-enabled jobs point to an inclusive model of employment, one that depends on continuous human intervention, a reality powerfully depicted in the film Humans in the Loop.
Much like the narrative shown in the film Humans in the Loop, AI systems may appear autonomous on the surface, but their intelligence is deeply shaped by invisible human labour. The film highlights how people; often at the margins – train, correct, and sustain intelligent systems. This idea closely mirrors India’s AI journey, where human oversight and contextual judgement are not optional add-ons, but foundational to building trustworthy AI at scale.
This human intervention becomes most visible when AI systems are deployed in real-world, high-impact public services.
The examples below demonstrate how AI systems that appear automated in practice rely heavily on Humans in the Loop; particularly in public services where accuracy, context, and accountability are critical.
Example 1: AI-powered traffic monitoring and enforcement in metro cities in India
Across metro cities, AI-enabled cameras are increasingly used to detect traffic violations and reduce road accidents. Violation of traffic rules such as signal jumping, speeding, phone usage, wrong-side driving, or not wearing seat belts, helmets are now being captured efficiently. The cameras automatically capture number plates and generate e-challans with minimal human intervention
For example, Hyderabad Traffic Police are deploying high-rise cameras for “Eagle Eye” to implement AI-powered traffic management systems; advanced AI-based cameras are already enhancing surveillance and traffic enforcement on the Outer Ring Road (ORR).
This may appear as a fully automated system, but in reality, humans play a critical role behind the scenes.
To make these systems accurately identify violations, AI models must be trained on thousands of real-world images and videos. This includes:
- What counts as “wrong-side driving” on a narrow Indian road
- How does the system distinguish between a slow-moving vehicle and one that has stopped due to congestion
- How does it recognise helmets, number plates, or lane markings under poor lighting or weather conditions
This foundational work is done by data annotators and quality reviewers, who:
- Label traffic footage frame by frame
- Identify violations, vehicles, and contextual factors
- Validate AI decisions before they are deployed at scale
These roles are being performed remotely and are well-suited for digitally trained youth and women, including those based in tier 2/3 towns and rural areas.
Example 2: AI-assisted medical imaging and diagnosis
Healthcare industry is one of the early adopters of AI to assist doctors in detecting abnormalities in X-rays, CT scans, and MRIs. AI has huge potential to help the doctors prioritise cases by identifying early signs of diseases such as tuberculosis, cancer, or retinal disorders, thus enhance the efficiency of public health systems.
While the AI output may look instantaneous, in reality, it is built on years of human effort.
Before an AI system can flag a tumour or abnormality, it must be trained on vast datasets of medical images where:
- Specific regions are marked and named
- Normal and abnormal patterns are clearly identified
- Edge cases and rare conditions are documented
This work is often carried out by medical data annotators and reviewers, working under strict protocols and supervision. These roles do not require individuals to be doctors.
However, they need:
- Training in medical imaging basics
- High attention to detail
- Understanding of structured guidelines
Such roles open up opportunities for life science graduates, women returning to work after a career break and trained healthcare workers supporting healthcare AI systems from the backend.
Humans in the loop (HITL): Potential for jobs and meeting inclusion goals at the base of the pyramid
“Every time an AI system flags a traffic violation or highlights a medical abnormality, it is standing on the invisible labour of humans in the loop—often trained, skilled, and working quietly from the base of the pyramid.”
The two examples show us how AI does not replace human work—it redistributes it.
From traffic enforcement to healthcare diagnostics, human judgement remains essential, especially in contexts as complex and diverse as India.
The examples also illustrate how humans can intervene at critical points in AI life cycle beginning with data annotation and labelling.
Hence, closing the loop of the AI model by plugging in human intelligence calls for policymakers and skilling institutions to reimagine AI jobs at the base of the pyramid by
- Recognizing these backend roles as part of the formal AI workforce
- Designing skilling programs that prepare individuals for such roles
- Ensuring these jobs are ethical, fairly paid, and offer progression
Now, are you wondering what kind of career pathways are emerging when someone works in the Traffic Management Systems? A Data Annotator can aspire to grow as Data Quality Analyst and later play a lead role in supporting projects like smart cities.
In healthcare, the Medical Image Annotator, after considerable work experience, can take up the role of Clinical Data Quality Reviewer and explore lead roles in AI Validation Specialist in the healthcare sector.
Why HITL jobs matter for India’s inclusion goals
India is uniquely positioned to leverage HITL work because of:
- Rapidly expanding digital connectivity
- A large pool of first-time job seekers
- Women seeking flexible and remote work options
- Linguistic and cultural diversity essential for training inclusive AI systems
These roles often do not require advanced degrees, but they do require:
- Digital literacy
- Language skills
- Attention to detail
- Human judgement
With the right skilling pathways, they can become formal, dignified digital livelihoods
AI jobs at the base of the pyramid with potential career pathways: Role and skills
1. Data Labeling / Annotation Associate
- The role involves tagging images, videos, text, and audio used to train AI models by identifying objects, intent, sentiment, and contextual meaning.
- The key skills required for this role are: Basic computer skills, reading and comprehension and familiarity with local languages
2. Speech and Language Data Associate
- This job role comprises tasks like recording, transcribing, and validating speech data. It can also include supporting Indian language and dialect datasets.
- The key skills required for this role are: Basic computer skills, language fluency, clear speech and basic keyboarding.
3. AI Content Reviewer / Moderator
- This role calls for reviewing AI outputs and user-generated content and flagging harmful, biased, or incorrect responses
- In terms of skills, besides computer skills, in involves critical thinking, cultural sensitivity and understanding of guidelines and following them carefully
4. Data Quality Analyst
- The role includes tasks like auditing labeled datasets and identifying error patterns and feedback loops
- Besides computer skills, in involves analytical critical thinking, cultural sensitivity and following quality assurance processes.
5. Human-in-the-Loop Trainer / Supervisor
- The key responsibilities of the trainer would include training new data workers, creating annotation guidelines and acting as an interface between technical teams and frontline workers
- In addition to computer skills and data labelling, the trainer needs to be strong in communication, process design and mentoring abilities
6. AI Bias and Ethics Reviewer
- This is an emerging role that includes evaluating datasets and AI outputs for bias and ensuring fairness across gender, language, and region
- Skills such as sound understanding of the domain (health, education, finance), ethical reasoning and structured evaluation are most essential for this role.
Together, these roles and skill pathways highlight how Humans-in-the-Loop are not just supporting AI systems, but actively shaping how AI is deployed, governed, and trusted across sectors.
We would also be glad to hear your perspective: What’s your AI story in education and skill development? Share your experience or express your interest in contributing by clicking the link here.
Building India’s AI Future with Humans in the Loop
India’s AI narrative is often defined by advanced models, computing power, and innovation milestones. Yet, beneath this visible layer lies a quieter but equally powerful transformation, where human intelligence, judgement, and contextual understanding continue to shape machine intelligence.
From traffic enforcement to medical diagnostics, Humans-in-the-Loop are not peripheral contributors; they are central to making AI reliable, ethical, and relevant for India’s diversity. These roles reaffirm that AI does not eliminate human work; it reshapes it, creating new digital livelihoods at the base of the pyramid.
For India, the real opportunity lies not only in building world-class AI systems, but in nurturing an inclusive AI workforce. By recognising, skilling, and formalising HITL roles, India can ensure that AI-led growth also advances social mobility, dignified employment, and sustainable career pathways, keeping humans firmly in the loop as partners in progress.
