When Hiring Algorithms Decide Who Gets Seen: Why Gladstone Samuel Believes AI in Indian HR Needs Stronger Governance

 

For more than three decades, Gladstone Leslie Samuel has worked in environments where a small oversight can have consequences far beyond the immediate problem. His professional journey has taken him through chemical process industries, industrial engineering, research and development, corporate leadership, risk management and governance. From working with organizations including ABB, Invensys, Baan, IDC and SPIC to serving as a board advisor, industry mentor and ESG professional, he has spent much of his career examining one fundamental question: how can organizations make better decisions while managing risk responsibly?

Today, that question is taking him into one of the most rapidly changing areas of corporate decision-making: artificial intelligence in recruitment. As Indian companies increasingly use AI-powered applicant tracking systems, screening tools, conversational bots and automated assessments to handle growing volumes of applications, Samuel believes there is a governance issue that businesses cannot afford to overlook. The technology may make hiring faster, but speed alone cannot determine whether the process is fair, inclusive or capable of recognizing talent outside conventional patterns.

His concern is not that organizations are using artificial intelligence. It is how they are using it.

For companies receiving thousands of applications for a single position, automated screening can appear almost indispensable. India produces a vast pool of graduates and professionals, while popular roles can attract enormous numbers of applicants within a matter of hours. Human recruiters cannot realistically examine every application with equal depth. AI therefore offers an obvious operational advantage by filtering large candidate pools and helping recruitment teams focus their attention.

But Samuel argues that organizations must be careful about what happens when historical hiring patterns become the foundation for automated decision-making. If a system learns from previous recruitment outcomes, it can potentially learn more than the characteristics associated with successful performance. It can also learn the preferences, limitations and structural biases that existed within the organization's historical talent pool.

That distinction is at the heart of his argument.

An organization may have historically hired a large proportion of graduates from elite institutions. It may have had fewer women in certain leadership or technical positions. It may have recruited predominantly from particular regions or professional networks. If those historical patterns become embedded into the data used by an automated screening system, the algorithm can begin treating yesterday's hiring patterns as a model for tomorrow's talent.

The result can be subtle. A candidate from a Tier-2 or Tier-3 institution may be overlooked despite having the skills required for the role. A professional returning to the workforce after a career break may find that the gap becomes an automated disadvantage. Candidates from less traditional socioeconomic or educational backgrounds may never reach the stage where a human recruiter can assess their actual potential.

In Samuel's view, this is not simply a diversity issue. It is a risk and governance issue.

His professional background helps explain why he approaches the subject differently from a conventional HR technology commentator. His experience spans chemical engineering, industrial safety, information technology, journalism and mass communication, while his professional qualifications include an MBA in Information Technology, a Diploma in Industrial Safety and a B.Tech in Chemical Engineering. He is also a Certified Independent Director accredited by the Ministry of Corporate Affairs, a Certified ESG Professional from the Confederation of Indian Industry and a Project Management Professional certified by the Project Management Institute.

Across these disciplines, one principle has remained consistent: systems that are not monitored can develop failure points that are difficult to detect until the consequences become visible.

Samuel's perspective on hiring AI is therefore rooted in the language of risk management. He believes organizations should stop treating AI recruitment platforms as ordinary software tools and start treating them as operational systems that require governance, testing and continuous oversight.

One of the most important questions, according to Samuel, is what data goes into the system in the first place. Seemingly neutral information can sometimes function as a proxy for characteristics that organizations should not be unintentionally using as barriers to opportunity. College categories, geographic indicators, age-related information or employment gaps can influence automated outcomes in ways that are not immediately obvious to recruiters.

For this reason, he advocates regular data sanitization audits at the input stage. The objective is to identify and remove high-risk proxy variables before they influence candidate scoring and screening decisions. In his view, responsible AI hiring begins before the algorithm produces its first recommendation. It begins with understanding what the algorithm has been taught to consider.

The second layer is examining what comes out of the system.

Samuel recommends using the 80% disparate impact rule as an important warning mechanism. Under this approach, organizations can compare selection rates between different candidate groups and investigate situations where an underrepresented group is selected at less than 80% of the rate of the highest-selected group. Rather than treating the threshold as an automatic conclusion of discrimination, it can function as an audit signal requiring further investigation, calibration and human review.

For him, the objective is not to make recruitment slower. It is to ensure that efficiency does not come at the expense of fairness.

The third requirement is greater transparency from HR technology providers. Organizations increasingly depend on third-party recruitment platforms, but purchasing software should not mean surrendering responsibility for understanding its consequences. Samuel believes companies should demand meaningful information about how these systems have been tested for bias, how they evaluate non-traditional career paths and how they treat employment gaps and unconventional credentials.

The fourth principle is perhaps the simplest: keep a human being in the loop.

AI can recommend. It can rank. It can identify patterns across thousands of applications. But Samuel believes final rejection decisions should not be left entirely to automated systems. Human recruiters need the opportunity to review cases where algorithmic outcomes could disproportionately affect particular groups or where a candidate's potential may not be captured by conventional data patterns.

His position comes from experience. Earlier in his career, while working in high-consequence chemical operations and industrial safety, Samuel learned that automated safety mechanisms could never eliminate the need for manual testing and human oversight. Systems can be powerful, but confidence in a system should never become a reason to stop checking whether it is functioning as intended.

That lesson has remained relevant throughout his career.

During post-acquisition integrations, he was also involved in situations where retaining talent became critical during significant organizational transitions. Achieving strong employee retention in such circumstances reinforced another belief: organizations are ultimately built by people, and trust, psychological safety and diverse perspectives cannot be reduced to a data point.

This is why Samuel views algorithmic fairness as much more than a corporate social responsibility initiative. For him, it belongs within the broader enterprise risk conversation. When an automated hiring system repeatedly excludes capable people because they do not match historical patterns, the company does not simply lose an opportunity to improve diversity. It potentially loses cognitive diversity, alternative perspectives and problem-solving capability.

The consequences can eventually reach the boardroom.

As an independent director and governance professional, Samuel believes boards will increasingly need to understand how artificial intelligence is being deployed inside organizations, including systems that may not traditionally appear on an enterprise risk agenda. Hiring technology can influence who enters an organization, which means its impact can extend into leadership pipelines, workforce composition, innovation capacity and corporate culture.

His message to HR leaders is therefore straightforward: AI should be an assistant, not an arbiter.

Instead of using artificial intelligence primarily as a mechanism for eliminating candidates as quickly as possible, organizations can use it to expand talent discovery and identify qualified people from broader and less traditional pools. Technology can help recruiters manage scale, but the responsibility for creating a fair and thoughtful hiring process cannot be outsourced entirely to an algorithm.

Samuel also believes diversity considerations need to become part of technology procurement itself. When organizations purchase HR technology, questions around bias testing, auditability and transparency should be considered alongside functionality, cost and implementation requirements. A system that saves recruiters hours but creates hidden exclusion risks may ultimately create a much larger organizational cost.

His governance philosophy also emphasizes that auditing cannot be a one-time exercise. AI systems operate in changing environments. Models process new information, organizations change their hiring requirements and candidate populations evolve. A system that appears balanced today may produce different outcomes later. For this reason, Samuel advocates periodic, including quarterly, disparate-impact reviews across automated hiring processes.

His broader professional work gives him a multidisciplinary platform from which to approach these questions. In addition to his corporate and industrial experience, he has served as an external board member, advisor, consulting partner, guest faculty member and industry mentor at institutions including Amity University, Dr. DY Patil University and LINC Education. His areas of expertise include stakeholder management, risk mitigation, corporate social responsibility and user documentation.

He has also contributed research on employee engagement and ESG audits for NGOs in India and has authored Corporate Governance in Indian Startups: Navigating Compliance and Growth, extending his engagement with governance beyond traditional corporate structures.

The debate around AI and recruitment is likely to become more important as Indian organizations accelerate their digital transformation. The question will no longer be whether companies use AI to recruit. It will increasingly be whether they can demonstrate that the systems they use are reliable, transparent and governed responsibly.

For Gladstone Samuel, the answer begins with a change in mindset.

Organizations should not ask only whether an AI recruitment system can process 50,000 applications in minutes. They should also ask who might disappear from that list before a human ever gets the opportunity to see them.

That question goes to the heart of responsible innovation.

Artificial intelligence can help organizations overcome the limitations of scale, but it should not reproduce the limitations of the past at a much greater speed. If companies want technology to create better workplaces, they must ensure that the systems making decisions about people are themselves subjected to the same discipline of accountability, testing and governance expected from other high-impact business systems.

For Samuel, the future of AI-enabled HR should therefore not be about choosing between technology and human judgment. It should be about combining the strengths of both.

The algorithm can process the scale.

The recruiter can understand the context.

And governance can make sure neither operates without accountability.

As artificial intelligence becomes increasingly embedded in Indian workplaces, that balance may determine whether automation becomes a genuine tool for expanding opportunity or simply a faster way of repeating yesterday's decisions.