Gartner report: Mitigate Bias From AI in Technology

The following is a summary of “Mitigate Bias From AI in Technology,” a report from Gartner. The full report is available for download here (For Gartner Subscribers)


Organizations are rapidly adopting AI in HR, while regulations are struggling to keep up. As part of their HR strategy, HR leaders must promote responsible AI in their applications by mitigating bias that poses risks to talent management, the employee experience, DEI and more.

Key Findings from the Gartner report

  • Fifty-three percent of HR leaders are concerned about potential bias and discrimination from AI. Bias in AI is unavoidable; however, HR leaders can establish best practices that mitigate this bias.
  • Thoroughly vetting HR technologies for bias requires an understanding of business processes. Broad overarching assessments can lead to missed sources of bias, or to inaction due to fear of getting it wrong. The organization should assess each use case individually.
  • AI regulations, including HR-specific measures relevant to bias, are gradually taking effect. Since many HR functions plan to buy AI capabilities built by vendors, HR leaders face the need to monitor technology providers for their regulatory compliance and ethical considerations.
  • HR leaders are positioned to take a leading role in advancing practices that bolster openness about the potential impacts of bias from AI applications, and 35% of HR leaders recently reported they expect to lead their organization’s enterprise wide AI ethics approach.

Gartner Recommendations

HR leaders responsible for technology strategy must:

  • Map possible sources and outputs of bias for each AI use case in HR to assist in flagging areas of risk and monitoring vendors for their commitment to responsible AI practices.
  • Require and evaluate bias mitigation from HR technology providers offering AI functionality by assessing criteria related to their data, algorithms, organizational context, regulation compliance and ethical considerations.
  • Promote transparency into the potential impacts of AI’s bias by collaborating with external and internal stakeholders to take decisive steps in protecting the organization, the future of work and society at large.

Gartner, Mitigate Bias From AI in HR Technology, By Helen Poitevin, 16 October 2023

GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved.

Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.

Webinar: and Citi Global Insights discuss how AI empowers HRs (Part 2 of 2) Co-founder and COO Isabelle Bichler-Eliasaf was recently invited to speak with Rob Garlick and Wenyan Fei from Citi Global Insights for a webinar entitled “Are Talent Intelligent Platforms the Solution to Reskilling and Upskilling?” 

Below is the second of a two-part blog series highlighting excerpts from their conversation. You can also view a recording of the full session here.

Isabelle Bichler Eliasaf, Rob Garlick, Wenyan Fei

You mentioned that represents fifth-generation AI. Can you share with us some unique features enabled by this latest generation of AI and what differentiates your platform versus other providers?

Yes, for sure. It’s really honing in on the skills taxonomies we’re creating; built one of the biggest data assets that exists today. We’re bringing a holistic solution from hire to retire, but the first thing you need to do is start with the skills architecture model. It uses the robustness of skills taxonomies to see details and skills, quickly analyzing all the roles of an organization. We’re providing a complete, dynamic catalog to our customer so they can have all their roles analyzed by skills–technical skills, soft skills, hard skills, and so forth. They have an understanding of the skills that are in demand, emerging skills, those that are declining, etc. This is the first step; really building the foundation in order to add on the next floor which is going to be talent acquisition or talent management and so forth. 

What we found is that a lot of organizations don’t do all that; the problem is that they operate in silos, without a unified skills language. They don’t share the same philosophy and the same strategy across teams. But cross-functional strategy should be shared because they’re going to use the same foundation of skills, the same framework. So what we’re doing at this stage is consolidating and synchronizing all teams in the same language. We’re taking the LMS skills framework and adapting it to the HCM and all the others used in an organization. That’s the first step, and now with that accurate, granular, robust skills data, I can inform talent acquisition; I can quickly create a job description based on the skills that I’ve analyzed for that role. Then I’m able to analyze the skills of your people; I can generate a heat map to understand what skills and organizational capabilities you have, and the skills gaps that represent weaknesses in specific departments. The skills data is also going to serve for learning and development to understand what the top skills are, and what the emerging skills are to fill skills gaps.

You say you’re a technology company built on data. Can you walk us through the kind of data that you ingest and how you train your model? It must be a huge amount to get to the point where you need to be in order to add value to people. Does this just get better and better as you are able to train it?

Exactly. It’s learning constantly. And the more data you collect, the more accurate you get so you can generate those predictions. What we’ve done throughout our first two years of inception is build a database by collecting from job boards, Indeed, all those different job boards, job descriptions, and so forth. Just on LinkedIn, you have 400 million active users that you can learn from; what career pathways they’ve pursued, and what skills they have for their position. We ingest all of this, plus pre-built taxonomies from Europe and Canada, and we continue to increase with market data. You can add even more from sites like Udemy; what are the learning trends, and what topics are in demand? All of this together feeds into one data asset. That is really integrated intelligence because we integrate with HCMs and older systems of record, not replacing them but just making them smarter with data. 

People are fearful about data privacy, algorithmic bias, bias in hiring, etc. How do we tackle this as a society? Is AI part of the solution or part of the problem? How are you guys doing that at in terms of responsible AI?

This is why I’m very passionate about responsible AI; because this is a problem and solution. AI is amazing. It’s like saying you’re not going to use electricity because there are risks. Of course you will, but you need to take some precautions. For us at, GDPR is what we’re embedding so that we’re compliant with the most stringent requirements. We’re also making sure all data we have is used with the consent of candidates; we know this because they chose to put their data online with all the information and informative disclosures involved. 

There’s also new legislation around telling you that you’re going to be assessed by a machine and allowing you to opt out: Local Law 144 here in New York City. I actually submitted public comments on the law along with others and one thing they hadn’t figured out is, if I’m a candidate opting out, am I able to be fully assessed? Am I actually discriminated against because I’m choosing that AI not be used? We had an amazing conference with Commissioner Keith Sonderling of the Equal Employment Opportunity and we were talking about that; his take on it is that the law didn’t change anything that wasn’t already in place. It didn’t do anything new. It’s still saying you shouldn’t discriminate based on gender, ethnicity, age, and all those different parameters–there’s another entity performing that assessment. I agree with him, but I think that the fear of AI makes us more aware and want to limit it. So now we need to balance between innovation and regulation.


Webinar: and Citi Global Insights discuss how AI empowers HRs (Part 1 of 2) Co-founder and COO Isabelle Bichler-Eliasaf was recently invited to speak with Rob Garlick and Wenyan Fei from Citi Global Insights for a webinar entitled “Are Talent Intelligent Platforms the Solution to Reskilling and Upskilling?” 

Below is the first of a two-part blog series highlighting excerpts from their conversation. You can also view a recording of the full session here.

Isabelle Bichler-Eliasaf, Rob Garlick, Wenyan Fei

What is a Talent Intelligence Platform? What is the technology that’s been transformed by AI to make upskilling possible?

Talent intelligence is really one of the biggest successes when it comes to AI for HR use cases. Talent Intelligence Platforms bring together all the relevant internal and external data sources into one useful tool to create a holistic view of a company that empowers decisions about talent or workforce planning. It’s really a merger between people analytics, market insights, market trends and workforce planning. 

If you’re looking at the system of records–HCM or ATS–AI makes them smarter. Why? Because it answers the most critical questions of recruiters: Who should they hire? What are the criteria based on skills, but not just basic skills? What is the experience required? This informs the recruiter as to who they’re looking for. With AI, they can take an omnichannel approach to sourcing candidates from two different pipelines: External sources, plus internal talent that may warrant a fresh look based on skills that are in demand now or that are emerging. HRs are able to hire much more efficiently by broadening the talent pipeline and looking at potential, not just unrelated qualifications and job titles. This is how talent intelligence is informing and empowering those HCM and ATS HR flows.

It’s my understanding that this is the same transformation technology that lies behind the large language models. If you don’t mind, maybe just contrast today’s tech with some of the legacy HR systems that create challenges for companies in terms of being able to extract data. 

What really has enabled us to bring increased efficiencies for retention and faster hiring and so forth are those models, of course, plus there’s been exponential growth–a leapfrog really–in AI This is because there’s much more data to train our models upon. In just the last two years, 90% of the current data was generated; that’s how big the computing power is now. It’s also been cheaper so we’re able to train better and more efficiently.

At the most recent HR tech conference in Las Vegas, analyst Josh Bersin recognized our work and categorized companies in AI into three types of companies. There are companies that are now adding AI into their flows, like adding a chatbot. Then there are companies such as SAP Oracle that have a lot of transactional data about candidates and employees and are now running some models to predict what would be a good fit, or what would be the next course an employee should take. Then there are the new types of companies such as ours that are built from the start on AI. We are the companies using neural networks. is using 1.8 billion data points and counting from all different sources to support HR flows; so it’s not just going to give an understanding of what you have in your company but also the ability to benchmark that against the market.