Responsible Recruiting in the Age of AI

Artificial intelligence is reshaping how companies hire, fire, and promote employees. The staffing industry has been quick to adopt AI-driven tools for efficiency, but are we fully considering the risks? While AI can streamline recruitment and improve decision-making, it also raises concerns about bias, transparency, legal exposure, and accountability.

One of my favorite quotes about applying technology to business processes comes from Bill Gates, who said in his 1996 book The Road Ahead: “Automation applied to an efficient operation will magnify the efficiency. Automation applied to an inefficient operation will magnify the inefficiency.” This is especially relevant when considering AI in hiring. I started my staffing career in 1997 as a Y2K project manager. In more than 20 years building technology to support business processes, I have yet to find a perfect one. We make poor choices in selection, we pass over perfect candidates, and we carry biases, both implicit and systemic. As we begin applying emerging technologies like AI to these processes, we have vastly expanded our ability to amplify the good or amplify the bad. But how will we know which is happening if we can't see into how these systems work or explain their decisions?

To explore this debate, I spoke with Hilke Schellmann, an investigative journalist and New York University assistant professor who takes a critical stance on AI in HR; Ben Eubanks, chief research officer at Lighthouse Research and Advisory, who advocates for a more balanced approach to AI adoption; and Fernando Rodriguez-Villa, a startup founder whose work in responsible AI caught my attention while evaluating dozens of tools as part of an AI program I developed in a prior role. His approach to explainability and third-party validation could serve as a strong model for responsible AI adoption in staffing.

The Legal Risks of AI in Hiring

AI adoption in hiring is increasingly subject to legal scrutiny, with concerns about bias, lack of transparency, and accountability leading to new lawsuits and regulations. Research into the legal risks of AI in hiring highlights key liability concerns and evolving governance frameworks that staffing firms must consider.

AI vendors can be held liable under antidiscrimination laws if their tools autonomously reject candidates (Mobley v. Workday Inc., 2024). Importantly, employers using AI tools are not absolved of responsibility. They must audit AI outputs and cannot rely solely on vendor claims.

Emerging legislation is reshaping how companies can use AI in hiring. The Colorado AI Act (2026) requires annual bias audits, disclosure to applicants, and transparency documentation. New York City Local Law 144 mandates independent AI audits and restricts AI use if bias is found. The Illinois AI Video Interview Act requires candidate consent and explanation of AI evaluation criteria.

To mitigate these legal risks, staffing firms should conduct regular bias audits, demand transparency from AI vendors, maintain human oversight in decision-making, and negotiate liability protections in vendor contracts. These evolving legal risks reinforce the need for responsible AI governance, aligning with insights from the panel of experts interviewed.

Industry Risks and Maturity of AI Governance

The Stanford University Institute for Human-Centered Artificial Intelligence AI Index Report 2024 provides critical insights into the state of AI governance. While privacy and security are top concerns globally, North American firms lag in addressing fairness concerns in AI systems. Alarmingly, only 17% of organizations in North America have fully operationalized more than half of surveyed risk mitigation measures for AI. The report also highlights a lack of standardized AI benchmarks, with vendors selectively reporting performance metrics, making fair evaluation difficult.

These findings underscore the urgent need for standardized AI governance practices in the staffing industry, where decisions directly impact people's livelihoods and careers.

The Case for AI Skepticism

In my interview with Hilke Schellmann, whose 2023 book The Algorithm: How AI Decides Who Gets Hired, Monitored, Promoted, and Fired, And Why We Need to Fight Back has become a touchstone for critical evaluation of AI in HR, she expressed deep concerns about the technology's implementation in hiring processes. Her two-year investigation into AI hiring tools revealed alarming patterns that staffing professionals should pay attention to.

Schellmann's research reveals that many AI-driven hiring tools inherit biases from historical hiring patterns. “You train the algorithm on résumés of people who were ‘successful’ in the job, and then it statistically finds patterns, often reinforcing bias,” she said. In her book, she documents how résumé parsers often act as hidden bias amplifiers, extracting data in ways that may favor or disqualify candidates based on arbitrary correlations rather than true merit.

Transparency is another major concern for Schellmann. “I wish companies would make their technologies available for testing. But most don't. Instead, we get marketing material saying their tools work and are bias-free,” she said in our interview, pointing out that most vendors refuse to share validation reports. This lack of independent validation is a central theme in The Algorithm, where she emphasizes that many AI-driven hiring algorithms have never been tested to see if they actually improve hiring outcomes.

Perhaps most troubling is Schellmann's reference to a study where 90% of surveyed executives admitted their AI systems rejected qualified candidates yet continued using them due to pressure for efficiencies. Her recommendation is clear: “The minimum requirement for any staffing firm should be a publicly available validation report.”

The Case for AI Adoption With Guardrails

When I spoke with Ben Eubanks, author of the 2022 book Artificial Intelligence for HR: Use AI to Support and Develop a Successful Workforce, he presented a more optimistic but still cautious view of AI in staffing. As he outlines in his book and confirmed in our interview, Eubanks advocates for a measured approach to AI adoption, emphasizing the importance of starting with lower-risk applications before tackling hiring decisions.

“These are low-risk areas…just a way for us to start getting some stuff off our plate before diving into more complex applications,” he explained when discussing his recommended implementation strategy. In Artificial Intelligence for HR, Eubanks presents a maturity model for AI adoption that begins with basic automation and progresses toward more complex decision-making systems. During our conversation, he shared examples of HR leaders who struggled with legal approval for AI tools, citing cases where compliance concerns halted implementation entirely.

On the topic of AI transparency, Eubanks is unequivocal both in his book and in person: “If I'm the head of talent acquisition and a vendor can't tell me how their AI works, I can't use them. I'm putting my team and company at risk.” He described a vendor tool that allowed recruiters to see exactly why candidates were ranked highly, an approach he believes should be standard practice in the industry.

Looking to the future, Eubanks predicts that while AI will continue to automate portions of the hiring process, human oversight will remain essential. “Companies that go all-in on AI without human oversight will eventually pull back. We're already seeing that in customer service AI, and hiring will be no different,” he said. Eubanks also highlighted an emerging challenge documented in his recent research: candidates themselves are increasingly using AI to optimize their applications, potentially creating AI-versus-AI scenarios that recruiters must navigate.

The Responsible AI Approach

Fernando Rodriguez-Villa approaches AI in staffing from a problem-first perspective, rather than chasing technology for its own sake. He explained during the conversation: “We started with a problem that we felt like AI could do a really effective job at addressing: identifying talent, matching it to jobs, and understanding the differences between candidates, even when there isn't a lot of information on them.” This focus on solving real problems helps avoid the pitfall of “AI for AI's sake,” where organizations implement technology without fully considering its impacts.

Transparency is central to Rodriguez-Villa's approach. His company AdeptID provides detailed documentation and resources on how its AI models work, believing this builds trust with clients and candidates alike. “Anyone should be wary of a website that says ‘we use AI’ but won't explain how,” he cautioned, seeing transparency not as a liability but as a competitive advantage in the market.

Rodriguez-Villa frames AI safeguards through an effective analogy: “It's not about slowing down as much as just having a steering wheel.” He compares current AI implementation to high-speed cars without seatbelts or antilock brakes, powerful but potentially dangerous without proper controls. This emphasis on explainability, he believes, is key to AI's long-term success in staffing.

The human element remains crucial in Rodriguez-Villa's vision. “If I am a recruiter and I see AI recommend a candidate, but I don't know why, I'm unlikely to trust it,” he noted. AI should provide clear reasoning for its recommendations to enhance, not replace, human decision-making in the hiring process.

Rodriguez-Villa also addresses the significant challenge of data quality in existing HR systems. “We wanted to build tech that worked for the majority of the workforce, even those not on LinkedIn or without traditional résumés,” he explained. AdeptID focuses on inference-based AI that can make decisions with limited information, addressing a common barrier to effective AI implementation in staffing.

A Path Forward: Implementing AI Responsibly

As staffing professionals navigate the evolving role of AI in hiring, structured governance is critical. The strategies discussed by the experts cited in this article — transparency, bias mitigation, third-party validation, human oversight, and legal risk management — can be codified into an AI playbook to guide adoption responsibly.

Components of an Effective AI Playbook:

1. Purpose and Scope Definition: Clear articulation of which hiring processes AI will and won't support
2. Ethical Principles: Documented standards for fairness, transparency, and accountability
3. Risk Assessment Framework: Methodology for evaluating potential bias and legal exposure
4. Validation Requirements: Standards for independent testing before deployment
5. Human Oversight Protocols: Defined touchpoints for human review of AI recommendations
6. Candidate Communication Guidelines: How to explain AI use to job seekers
7. Continuous Monitoring Plan: Process for ongoing evaluation of AI performance and bias
8. Incident Response Procedure: Steps to take if bias or errors are detected

A strong AI playbook helps align use with company goals, ensure ethical practices, and track impact. It sets clear policies, outlines key use cases, and establishes risk controls to guide responsible adoption.

Measuring Success

Organizations should establish clear metrics to evaluate their AI implementation: reduction in time-to-hire without sacrificing quality, improvement in diversity of candidate pools, recruiter satisfaction and trust in AI recommendations, candidate experience ratings, reduction in bias incidents, and quality of hire metrics compared to pre-AI baselines.

Balancing Innovation and Responsibility

The perspectives of Schellmann, Eubanks, and Rodriguez-Villa offer a comprehensive framework for approaching AI in staffing. Schellmann's skepticism reminds us to question vendor claims and demand evidence. Eubanks' pragmatism provides a roadmap for gradual, risk-managed adoption. Rodriguez-Villa's responsible implementation approach demonstrates how transparency and human-centered design can build trust.

As AI continues to transform the staffing industry, the organizations that will succeed are those that view AI not as a replacement for human judgment but as a tool to enhance it — one that requires careful implementation, ongoing oversight, and a commitment to ethical principles. By embracing both innovation and responsibility, staffing firms can harness AI's power while protecting the diverse talent they seek to place.

A Note About AI Use: This article was developed using a combination of expert interviews, personal analysis, and AI-assisted research tools. The core insights are drawn from direct interviews with Hilke Schellmann, Ben Eubanks, and Fernando Rodriguez-Villa, as well as firsthand evaluation of AI hiring technologies. Perplexity AI was used for legal research on AI regulations and risk mitigation strategies. OpenAI ChatGPT assisted in structuring and refining the article's key arguments. Anthropic Claude provided critical feedback on content clarity and alignment with industry best practices.