Skip to Content

Report Summary: Understanding AI in HR by Josh Bersin


Everybody’s talking about the ways in which AI is both revolutionizing and disrupting human resources – with a mix of fear and excitement – and it can be hard to keep up with the rapid pace of change. HR and L&D adviser and expert Josh Bersin distills this complex topic into applicable insights, as he fills you in on the ways AI will transform the industry. Whether you’re interested in using AI to mitigate bias or in harnessing AI’s potential to predict job performance, it’s crucial that you identify the tools that work best for you and gain a basic understanding of the new generation of AI solutions on the market, explains Bersin.


  • Solve your human resources (HR) problems with three different types of AI solutions.
  • Second-generation AI systems spot meaningful patterns to predict performance.
  • Use AI tools to mitigate bias, ensuring diverse recruiting and pay equity.

Report Summary: Understanding AI in HR by Josh Bersin


Solve your human resources (HR) problems with three different types of AI solutions.

Artificial intelligence (AI) is transforming the world of HR, enabling recruiters and companies to automate tasks ranging from predicting performance to writing job descriptions. Aim to familiarize yourself with the following three categories of AI solutions:

  • Emerging AI – Vendors leveraging data management and analytics to create dashboards, are offering AI systems that you “add on” to your existing HR platforms, which generate data-driven reports and predictive models. Emerging AI systems can encompass applicant tracking systems (ATSs), which automatically generate job descriptions and customized scripts for interviews. These solutions don’t use large language models (LLMs), but sometimes use generative AI, as well as statistics and principles gleaned from pure AI.
  • First-generation AI – Vendors using machine learning (ML) and AI models typically work with AI engineers to comb through large data sets to provide intelligent recommendations via “built-in” solutions. For example, first-generation AI solutions may match candidates to organizations or recommend users targeted upskilling courses based on their skills and job role. These systems don’t rely on neural networks, and lack the power of deep neural networks and GPT-4. Examples of first-generation AI include enterprise research planning (ERP) vendors, such as SAP, Oracle and Workday.
  • Second-generation AI – These “built-on” solutions refer to entire platforms designed specifically for AI. Data-centric companies are turning to vendors such as Gloat, Eightfold AI and SeekOut to glean better insights and to better manage massive amounts of data. These platforms are powerful because they harness natural language processing (NLP), deep learning, vector databases and LLMs, drawing both from your data and that of hundreds of millions of external profiles to deliver insights. Consider building your own second-generation system with cloud service offerings from Microsoft, Google or OpenAI.

Second-generation AI systems spot meaningful patterns to predict performance.

HR firms that want to expand their reach will get the most value out of second-generation AI, as these platforms can create their own predictive models, quickly identifying ways to achieve the goals you program them with. For example, to understand why employees in a particular region are performing poorly, a second-generation AI system could sift through millions of employee profiles to find factors correlating with poor performance. Using neural networks, the AI “learns” and improves itself when it accesses more data, enabling it to “get smarter and smarter” in its approach to finding solutions. For example, drawing from a massive employee and job candidate database, an AI can identify numerous patterns – much like a human brain does but at a large scale – looking at factors such as job histories, personal connections and educational history, to identify those shared by high performers.

“The exciting thing about AI is that it doesn’t ‘guess’ what’s going to work. It simply creates a model.”

Advanced AI systems can help recruiters and organizations spot candidates with their desired skills, assessing data sources ranging from candidates’ certifications to GitHub code repositories. Second-generation systems have more capacity to avoid semantic confusion when it comes to predicting skills than emerging and first-generation models. For example, a second-generation system wouldn’t mistake a barista with the word “Java” on their résumé or someone living in the country Java for a developer who knows the programming language JavaScript, as it remembers the relationships between words. These systems can also spot skills adjacencies, identifying related terms to predict desired skills. For example, one AI system found a link between leadership skills and “soft skills,” ultimately guiding the client toward hiring those who modeled certain interpersonal behaviors and values.

Use AI tools to mitigate bias, ensuring diverse recruiting and pay equity.

It’s crucial to ensure your AI models train on unbiased data, especially as some US states now mandate diverse hiring, pay and promotion practices. Vendors such as Eightfold AI, iCIMS and SeekOut offer clients proof that their models are “unbiased,” but eliminating bias is an ongoing process for all involved. Ultimately, the bigger data set you use, the less your company’s specific biases will be reflected in your AI system’s insights and recommendations. Companies that have failed to correct bias have paid millions of dollars to employees in lawsuits. Google, for example, paid employees more than $118 million when they proved the company had practiced gender discrimination in promotion and hiring.

“As you can imagine, if you train a system on biased data you get biased results. That’s why a small AI model based on your company’s data may lead to problems – every bias embedded in your company will be amplified.”

AI can be leveraged as a tool to help ensure pay equity, as you can train models to counteract your own organization’s tendency toward bias in promotion, pay and advancement. Simply add your company’s payment data to your AI system, prompting it to make “predictions for pay,” to note which groups of people are overpaid and underpaid, while spotting patterns reflecting bias. When you mitigate bias and choose the best AI system for your specific needs, AI can help you accomplish a variety of tasks, including diverse recruiting, capability development and career planning.

About the Author

Josh Bersin is the founder of Bersin & Associates, a company that provides research and advisory services in areas such as leadership, HR and L&D.

Alex Lim is a certified book reviewer and editor with over 10 years of experience in the publishing industry. He has reviewed hundreds of books for reputable magazines and websites, such as The New York Times, The Guardian, and Goodreads. Alex has a master’s degree in comparative literature from Harvard University and a PhD in literary criticism from Oxford University. He is also the author of several acclaimed books on literary theory and analysis, such as The Art of Reading and How to Write a Book Review. Alex lives in London, England with his wife and two children. You can contact him at [email protected] or follow him on Website | Twitter | Facebook

    Ads Blocker Image Powered by Code Help Pro

    Your Support Matters...

    We run an independent site that is committed to delivering valuable content, but it comes with its challenges. Many of our readers use ad blockers, causing our advertising revenue to decline. Unlike some websites, we have not implemented paywalls to restrict access. Your support can make a significant difference. If you find this website useful and choose to support us, it would greatly secure our future. We appreciate your help. If you are currently using an ad blocker, please consider disabling it for our site. Thank you for your understanding and support.