HR and artificial intelligence: What’s next?
Artificial intelligence has already more than proved its ability to support HR decision-making, whether it be to win, retain or develop talent. But what more can we expect moving forward? We ask the expert: Cyril Le Mat, Director of Data Science at Cornerstone OnDemand.
First of all, let’s clarify what we mean when we talk about “artificial intelligence”. Contrary to popular belief, it is not about making and programming machines that mimic human intelligence. Artificial intelligence simply refers to the intelligence of machines, as opposed to that of humans. Specifically, it refers to their ability to analyse their environment and to react and adapt to it as machines. The aim is not to replace humans, but rather to support them by executing tasks that machines can perform better and faster.
AI research in the HR space focuses on the analysis and cross-referencing of three types of information: profiles, positions and training content. “One of the big challenges with HR is the fact that the data to be analysed is mostly unstructured text containing terms that are often very subjective,” explains Cyril Le Mat. “Legal texts are objective and leave little room for interpretation. This is not the case in HR. A CV is difficult to analyse because the concepts behind the words vary greatly depending on the individual. Given that HR data is intrinsically complex, most systems have avoided using this information in the past.”
AI research in the field of HR has therefore largely focused on identifying personal competencies, the nature of the connections cross-linking the various skills, the definition of job profiles, and the cross-referencing of all this data.
How AI is supporting HR today
AI tools already support – at the very least – the following basic functionality:
- Detecting and extracting an individual’s skills from their documents on file (e.g. CV, cover letter) in order to create a skills graph, for example;
- Translating a job profile into 15-20 skills instead of 2 or 300 tasks to be performed;
- Matching offers and profiles.
In what kind of HR context are these technologies used? There are three main ones:
- Recruitment: AI helps to formulate the skills content of job offers in a relevant and effective way; to identify candidates’ skills; to perform an initial selection of candidates if necessary; and to support the candidate’s journey.
- Talent marketplace: The same technologies are used to enable internal mobility and manage career paths. “AI uses auto-detection to recommend content and relevant opportunities to employees based on the skills that the tool helps to identify from their CV and any other document in their HR file,” explains Cyril Le Mat.
- The LXP (Learning Experience Platform): AI not only helps employees identify their existing skills but also to identify skill gaps that need to be bridged to achieve a professional goal. It can then suggest the training best suited to supporting this journey.
All three of these applications are already available on the market, albeit at different levels of development and performance.
The future evolution of AI
AI is not a single static technology on a rapid trajectory towards market maturity and widespread standardisation. By its very nature, artificial intelligence is in a constant state of flux. As machines get better at learning, we can – over the coming years – expect them to get better at doing what they already do today. AI applications for HR have already shown their efficacy and are set to become even more powerful. “Everyone has their idiosyncrasies, their own ways of describing things,” continues Cyril Le Mat. “AI will get better at understanding how each individual expresses themself so they can assess that person’s skills more accurately.” The result will be increasingly tailored responses and services. “This will be a core market trend for the next ten years.
According to Cyril Le Mat, however, AI’s contribution to HR will not stop there. “The organisation-wide visualisation of skills will make it possible to support HR decision-making and strategic workplace planning. The idea is to use this “skills” data provided by employees to produce visual maps that will help inform decisions by company management. The aim is to evolve HR beyond a reactive function (we have a vacancy we need to fill) to a strategic role capable of performing skills gap analysis and thus identifying future training and recruitment needs to achieve the company’s goals, as well as being able to identify development opportunities from pockets of skills, and so on. These visualisations allow HRMs to go to the decision-makers and business leaders with a HR vision for the skills and opportunities available to the company.”
Simple in theory, using AI to support strategic decision-making still requires some fine-tuning and preparation for mass roll-out. The technologies are already well advanced, however. “We have around 50,000 defined skills. You could find 5,000 or so in any given company. The challenge is to represent all these skills by mapping them. The 1,000 most important skills will be pooled together according to similarity. Visualisation allows you to zoom in and view each skill in its context beside other similar skills, but also to ask questions and compare two entities. We will see a lot more products of this type being developed to aid strategic global decision-making within organisations.”
Limitations of AI
There is huge potential for the development of artificial intelligence in qualitative and quantitative terms, whether in HR or other applications. At the same time, the technology has its limits. “No AI system could claim a detection rate of 99.99% with this type of data,”says Cyril Le Mat. “The challenge is to improve gradually from 95% to 96%, 97%, and so on.” There are at least two things that AI will not be able to do in a HR context though:
- Identify soft skills with absolute accuracy. “These technologies are very good at recognising anything related to hard skills. When it comes to soft skills, however, it’s hard to get reliable results that really make sense. And we always have the risk of discrimination in HR. We cannot make strong claims in this area. Companies that claim to identify soft skills via AI either stop doing it or are not really doing it.”
- Replace humans in decision-making. This is not what AI is all about. According to Cyril Le Mat, its purpose is to “automate tasks that can be automated, leaving humans to focus on tasks with higher added value.” In short, AI suggests, but humans choose and decide. “It is important to be aware of the limits of your technology.”
AI’s contribution to HR lies above all in this complementary relationship with humans. “We can say with relative certainty that AI will never replace human decision-making, but it does free up the time spent looking for information and analysing documents to open up opportunities for employees. I was surprised by the extent of the positive feedback from employees and HR managers alike. Tools that incorporate AI are changing the relationship between employees and HR. For example, employees no longer reach out to their HR department for information, but rather to say ‘I’ve seen this opportunity, can we talk about it?’” Contrary to initial fears, artificial intelligence does not dehumanise workplace relationships. (https://www.cornerstoneondemand.com/fr/resources/article/l-intelligence-artificielle-va-humaniser-le-travail-et-non-le-remplacer-fr/) Instead, it benefits everyone involved by giving them more autonomy and greater capacity to take the initiative.
Want to keep learning? Explore our products, customer stories, and the latest industry insights.
Empower HR with the latest and greatest AI technology
AI is here to disrupt the traditional and empower the innovative.
A pillar of progress: AI for HR depends on good data
In the ever-evolving landscape of Human Resources (HR) technology, Artificial Intelligence (AI) has emerged as a game-changer. AI-driven tools promise to streamline HR processes, enhance decision-making and revolutionize the employee experience. However, the efficacy of these AI solutions is only partially dependent on the sophistication of algorithms or the power of computational hardware. In fact, a large part of the equation hinges on a fundamental foundation — good data.
How AI powers skills matching and ontology
Have you ever played match maker? It's no easy feat, and in the world of work, matching an employee's skillset is often very challenging. However, technology has a crucial role to play in streamlining this. Leveraging AI to build skills ontologies will create more accurate and dynamic databases, driving the future of AI-powered skill matching.