Enhancing Fairness and Efficiency in Human Resources With AI

Dataiku Product, Scaling AI Lynn Heidmann


At its best, the impact of AI can help make Human Resources (HR) more effective and more fair. It can make hiring and personnel policies more efficient by working at speeds unfathomable to a team of people, let alone an individual. It can work independent of the myriad human biases that infect personnel decisions. According to a McKinsey report, GenAI deployments are bringing efficiency and new insights to HR functions and are estimated to enhance HR productivity by up to 30%, with examples demonstrating reduced annual budgets and improved efficiency.

However, as is the case with most technological innovations, AI (including GenAI) presents a number of potential pitfalls. Without proper safeguards, they can infringe on employee privacy or create a “big brother” ethos that runs the risk of prompting resentment from workers. In some cases, algorithms designed by humans can serve to further entrench biases, rather than eliminate them. 

Finally, regardless of its use in HR processes, GenAI will be an important theme for hiring managers simply because of the high demand among employers for employees trained in AI. 

A More Human-Centered HR Workforce

GenAI is increasingly playing a significant role in screening applications. This is particularly important for large employers who may receive thousands of applications for a single position. A 2023 analysis by McKinsey estimated that GenAI and other technologies have the potential to automate work activities and save up to 70%of employees’ time. 

But that doesn’t mean AI will replace humans, even in HR. Instead, AI is slated to empower remaining workers to dedicate more time to tasks that require a human touch. In many smaller organizations, HR personnel have long had to dedicate substantial time to bureaucratic tasks that take time away from the more important work of closely evaluating job applicants and considering ways to better-engage the workforce. 

Are Employers Ready? 

Organizations Voice Doubts 

The potential impact of AI on HR is undeniable and employers are clearly eager to harness the advantages. AI-powered screening tools, as found by a study conducted by Talent Board and Phenom, have been shown to reduce the time spent on résumé reviewing by up to 75%, highlighting the transformative power of AI in streamlining talent acquisition processes. Furthermore, Phenom's research reveals that AI's role in interview scheduling has been noteworthy, with 80% of organizations reporting time savings of 36% compared to manual scheduling methods. These statistics underscore the widespread adoption and tangible benefits of AI in revolutionizing talent acquisition practices.

What Will It Take?

With all of that doubt, what will it take for HR departments to get ahead in the age of AI? Spoiler alert: There isn’t a magic bullet, one-size-fits-all solution — it will take a combination of organizational change in people, processes, and technology to empower all HR staff to use data in their day-to-day work:

People need to trust in machine learning (ML)-based solutions. Part of this trust comes from education and transparency; that is, ensuring that processes are white box enough for everyone to understand. Choosing the right technology and instituting training programs can help individuals better manage and embrace change. Learn more about how one company, GE Aviation, developed extensive training programs for staff to gain skills and knowledge about working with data to empower even non-data experts.

Processes need to be developed to encourage the more prolific use of data among and across teams. For example, HR staff should be able to work collaboratively with both data experts and IT teams (whether or not those roles are embedded into the HR team itself). 

Technology should round out people and process strategies by providing a governable framework from which to support both large-scale, operationalized ML models and AI projects as well as smaller scale data exploration by HR staff. Choosing technology that is flexible (not locking the company into one solution or approach), innovative, and supports Responsible AI is also critical.

High Demand

The demand from employers for workers who can work in AI far outpaces the supply of qualified employees. Between 2023 and 2024, the demand for GenAI-related talent in the U.S. rose more than 30-fold, Hiring Lab found. Dice also reported that 14% of all job listings in February 2024 referenced skills related to AI or ML.

That helps explain why more than 700,000 AI positions with U.S. employers remain vacant because companies are unable to find qualified candidates. Simply put, the technology has developed more quickly than universities and other training programs can respond.

Finally, staffing a business with the necessary AI skills will require not just hiring, but significant retraining. Employers cannot depend on the educational establishment to readjust quickly enough to provide qualified job candidates. HR will play a crucial role in connecting employees with training opportunities that will provide them greater job security and the company with greater value. 

Potential Pitfalls 

Privacy

The use of AI in HR presents a number of thorny ethical questions for individual employers and governmental bodies. As is the case with many other issues surrounding technology and data privacy, it is likely that rules will differ significantly between jurisdictions, but there are two major global concerns:

It’s one thing for employers to use algorithms to analyze data that job applicants or employees provide them, such as on resumes. The public will likely feel less comfortable, however, with employers using AI to dig deep into a candidate or employee’s personal life, such as analyzing years worth of social media behavior or publicly accessible legal records: marriages, divorces, land deeds, criminal offenses. 

Local governments throughout the U.S. have already taken steps to bar employers from considering certain factors in the hiring process. Many states and cities prohibit employers from asking applicants to disclose whether they have a criminal conviction on their application. Many others have barred employers from asking applicants to disclose social media usernames or passwords.

However, the increasing sophistication of algorithms may allow organizations to evade these restrictions by unleashing algorithms that scour the internet and analyze information on applicants that would have taken a team of HR personnel hours, if not days or weeks, to dig up on their own. Such tactics may prompt calls for new laws that strictly regulate just how deeply employers are allowed to dig in the application process.  

In some cases, privacy concerns will translate into new laws. In other cases, however, they may result in something even more threatening: backlash from consumers and low employee morale. 

Bias

In some cases, algorithms may not break down biases, but reinforce them. Amazon suspended its work on a hiring algorithm after it revealed strong evidence of bias against female applicants. The program selected candidates based on patterns it observed among top employees or candidates. Since tech is predominantly male, the program viewed female candidates as a deviation from the pattern and penalized them

Bringing the Human Touch to AI in HR

Just like many other technological innovations in recent decades –– from the fax machine to the internet –– AI will bring major benefits to HR. Just like those other technologies, it will significantly broaden the reach of recruiters, allowing businesses to more effectively scour the Earth for high-quality job candidates. 

Similarly, AI offers the chance for employers to better understand their human capital. Algorithmic programs can uncover qualities in the workforce that would otherwise go undetected by a team of busy HR leaders. Algorithms that can quickly assess mountains of employee-generated data –– hours worked, projects completed, disciplinary actions –– will help employers discover strengths and weaknesses in their workforce.

However, as they move towards expanding their use of AI, employers must recognize that algorithms are not immune from human error. After all, they are designed by humans and therefore inherit many of the same assumptions or flaws of their human creators. 

Furthermore, it is critical that employers be wary of privacy implications and communicate clearly with workers about what data is being analyzed and what they can expect to be private. For an effective approach, employee buy-in is a must. They must understand that AI is not merely a means to replace workers, but a way to enhance the work experience for everybody at the organization. 

Next Steps

The first step to leveraging AI in HR is to start fostering a data-driven approach and understanding of what AI is and how it can help — not replace — HR efforts:

  • Empower HR staff by teaching ML basics. Get started with this blog series for non-technical readers instead: getting started, building the model, and interpreting it.
  • Ensure a deeper understanding at both the individual and management level of what it means to do Responsible AI. That includes ML interpretability, data privacy, and ethics.
  • Start exploring processes for leveraging data within the HR team. That might mean hiring a data expert to sit in the business unit or empowering those who already have deep knowledge of HR systems to start using data.
  • Choose the right technology for the AI journey — especially in the age of GenAI. That might mean considering a data science, machine learning, and AI platform that aids with governance and compliance plus is accessible to all profiles, from HR professional to advanced data scientist, to work together toward advancing the company’s AI aspirations.

You May Also Like

Maximizing Text Generation Techniques

Read More

Looking Ahead: AI Hurdles IT Leaders Need to Overcome in 2025

Read More

No-Code ML and GenAI With Dataiku and Fabric

Read More

Unpacking 3 of the Biggest Controversies in AI Today

Read More