Can AI really replace HR professionals... and how can I inform myself about AI?

Steve Bridger

| 0 Posts

Community Manager

8 Mar, 2024 12:32

Immediately after his closing conference keynote about the impact of technology and AI in particular on the profession,  was asked some questions. While David responded at the time, we thought it would be interesting to explore the topic a little more here - especially for those who weren't able to make it on the day. 

As an HR technology novice - what can I do to gain more knowledge?

Do you think AI within people practice will be used to help remove the monotony tasks. I assume HR maybe more focused on strategic/innovative/non-routine decisions that AI possibly cannot do - do you agree with this?

Can AI replace HR professionals?

I'll ask David to respond below.

As an aside, members are also discussing these themes over in the  AI in HR forum.

  • Excellent question! A few thoughts on that final question below

    Whenever we are talking about the capability of technology it's worth thinking in specific timeframes. The 'future' is a very broad timeframe. Zip back to 5 years ago and we are before the pandemic. 10 years ago Zoom hadn't been released yet and go back 30 years and you'd find computers weren't universally available in offices. You might be allocated a word processor if your job needed it.  

    As time passes some of the basic assumptions of how we work and live will always change. More time means more change.

    As technology has evolved it has been able to do more of the things that we can already do - and some of the things that we can't. It's superhuman in some respects and far less than human in others.

    It's allowed us to do things that previously we wouldn't have been able to within the profession - imagine trying understand to the demographics of a large organisation before records were digitised. Or perhaps attempting to doing payroll for 1500 people, but using just a pen, paper and cash... It's as much about what technology can do for us as it is about what it can do standalone. That's why we need to see the risks and opportunities in parallel.

    If you are worried (or excited...) about AI replacing HR professionals then it's worth thinking about a fixed time frame - for instance the next 5 years - and writing a list of what you couldn't expect technology to do as well as a professional within that time frame. Then revisit that list on a regular basis and see if your assumptions should change. 

    Perhaps start with your current job description, but make sure you augment the basic list of tasks with the understanding of what you really do within them. We know that processing people is different to really supporting people through a process - for instance. If you have your career plan in place then you could repeat the process with your next intended role too. 

    Examples of work you might not expect technology to be able to replicate/deliver as well might be 
    i) work that relies on relationships or trust 
    ii) work where people make a difference through impact/influence 
    iii) work where regular judgement calls are needed that rely on expert knowledge being blended with the your understanding and judgement

    It's a good exercise to think about the issue in a more specific way and also understanding the areas where you will be able to add most value. If you do that and make use of our resources you'll be in the best possible place to position your career to surf the wave of some of these changes. 

    Let me know your thoughts...
  • In reply to David D'Souza:

    i think AI will be used to automate some of the more transactional elements of HR, it will also be used for things like transcription of interviews etc. I am sure on the TA side it will become more and more prevalent in the initial acquisition and screening of applicants.

    However the Human bit of Human Resources pretty much dictates that all the time people are doing jobs and not AI people will have to be involved in managing the human intricacies of that employment
  • In reply to Ian:

    It will absolutely be used for some of those areas that you've described. I think there's less of a risk in note taking/transcription than in the choices that sit within selection - but equally we know there's an opportunity for improvement in outcomes there too
  • My husband's firm are developing a HR Bot to deal with first level HR queries. It's not yet fully operational, but I expect, a direction of travel. Will we get to a position they can offer nuanced advice? Maybe, but that is a way off yet. But direct people to policies, answer that question about how many days holiday do I have again? Totally.
  • In reply to Gemma Dale:

    There's levels to it - surface broad policy, contextualise policy, give advice, give contextualised guidance. I think we'll see a lot of providers suggesting they are doing the more complex pieces when it's the first grouping... Say 'hi' to your husband too!
  • Thinking about the things I've done this morning, pretty much all of it fits into the 'human' and nuanced category. But when I think about the things that make up an HR career and role, what worries me is that the reason I feel confident doing the more complex things now is because I've spent a long time doing things that I might reasonably expect to be done by AI in future.

    How do we expect the HRDs of the future to develop their skills and understanding, if they don't get the kind of experience that builds things gradually? I hear these same concerns expressed in other professions too, and I think this is something we need to focus on as HR professionals - for ourselves and for the other professionals we work with.

    Scary and exciting, but definitely one to try and engage with.

  • I can add few bits too, it is unlikely to fully replace HR professionals, contrary to that Artificial Intelligence has the potential to augment many aspects of HR functions. For example Organisational Culture and Values. AI algorithms may lack the contextual understanding and cultural sensitivity to manage diversity, equity, and inclusion initiatives. One good thing AI would do is, freeing up time for HR professionals to focus more HR strategic initiatives, such as decision-making, and value-added activities for greater HR competitive advantage
  • In reply to Gemma Dale:

    The concept of the HR Bot to deal with first level HR queries is an interesting one and I wonder how employees would take to it?

    Lets consider that all of us as customers of many retailers/utility providers are already experiencing dealing with these chat bots on a daily basis whenever we make an attempt to contact any kind of customer service department and navigating them is quite frankly a pain! I've lost count of the number of times I've finally gotten through to an actual person after doing virtual battle with the chat bot who seems to be put in the way to prevent us from actually getting to customer services and been more annoyed/frustrated than I was when I first made contact.
    It's the same with the voice technology if you call a company - its like dealing with the Riddler trying to make sure you word your query correctly to get to the right person.

    I can see how it would work for things like checking holiday balance - enter your employee number and hire date and we'll pull it from the system. But if as an employee I have a question but first have to jump through the chat bots hoops to get to speak to my HR person then I'm likely to be more stressed/worried/upset by the time I get to them. Which then for the HR person on the receiving end is potentially going to make helping said employee more difficult.
  • Hi Steve, I do a lot of tech work and have a few thoughts to add to some of these questions.

    what can I do to gain more knowledge? I'd recommend the following:
    * dive into the space, remain curious and sceptical, be open to learning and mix in different spaces. I've found a lot of learning and synergy my mixing with tech colleagues, going into tech spaces, mixing with other disciplines like business analysts and going on courses or formal training in the tech space.
    *read and read beyond blog, Linked In posts and vox pop. There are some great books, articles and learning in this space.
    *follow credible tech voices, as you gain more knowledge you will find even more people to follow and know how to spot those who might be slightly over egging or under cooking the impact of tech.
    *don't be afraid of the lingo. Tech speak is like HR speak, shorthand that can sometimes provide an accessibility barrier. Plough through it. There really is no such thing as a silly question.
    * tech is broad and deep. I don't think it's possible to know it all so it's important to keep learning to deepen our knowledge.

    Personally, I believe this is a space full of differing levels of impact and experience. For some companies they are still running things off spreadsheets and don't even have an HRIS/HRM/HCM and good data analytics let alone are ready to explore looking automation, early AI people solutions, chat bots etc. Other companies have invested heavily and might be looking at different challenges and then there is everything in-between.

    AI is a long way off replacing humans in many areas of the world of work and I find it very interesting what tasks are the focus of tech attention.

    AI is a product so there is an element where it will look for easy roots to market and those that can be commercialised. It will definitely evolve and impact our work, in some ways we expect and others we don't. It will create opportunities, challenges and issues. I'm particularly interested in the areas of bias, ethics, security and the impact of tech on culture and behaviour.

    It's going to be an interesting ride and the key is to keep evolving too so we can survive the rollercoaster.
  • Recommended reading: "Godel, Escher, Bach: An Eternal Golden Braid" by Douglas Hofstadter.

    Warning: it is a dense, hefty and very complex work and its title is no indication of its content which touches only incidentally on the three masters it names. It is full of wordplay, metaphor and surrealism. But it is also the most influential and prophetic book on the subject of artificial intelligence ever written.

    Essentially, Hofstadter makes a number of predictions about the nature of human intelligence and how it might arise artificially and why it basically can't be built intentionally but can only arise as an emergent property of a sufficiently complex system. This, of course, is about what has now been rebranded as "Artificial General Intelligence".

    What is being built now is not "intelligence" in any real sense, but algorithmically-generating systems. What makes modern versions interesting, though, is that the algorithms are themselves being generated algorithmically and iteratively to the point that we no longer really understand how they work (something else that Hofstadter predicted).

    The result is that, whilst Generative Systems can create the impression of understanding their dialogue and providing useful advice, *they do not*. They generate probabilistic outcomes based on things that their training data has told them look like correct answers to their questions. Consequently, they are not only not infallible, but they are fallible in ways that are entirely inhuman.

    You will have seen this in bad AI art that doesn't understand how fingers works. They have no true appreciation for the illusion of two-dimensional art. They are just reproducing something that resembles the examples in their training data that might correlate with what they have been asked to produce.

    What all of this means for AI in HR is that, yes, you can definitely use these new Generative Systems to help produce letters, compose policies, answer routine questions and extract some of the drudgery from our profession. But you cannot let them do these things unsupervised. And understanding how to supervise the work of a Generative System is a completely new form of management that most people will not be trained in or understand. We must never pretend that the work of a GS is the work of a human because, when they mess up, the lie will be instantly exposed because the mistakes they make will not be human mistakes because they don't arise from anything resembling actual "intelligence".

    We must always provide GS with a human back-up and never release anything created by a GS into the wild without first having it reviewed by a human.

    Generative Systems are getting better and better, of course, at an astonishing rate. But this improvement comes at a cost, which is that they require greater and greater amounts of training data and greater quantities of processing power to contain and process this data. This means that the best GSs can only exist on enormous server banks that must, by definition, be owned and controlled by someone human. And humans who own and control server banks of this size have agendas that will inform the training data that goes in and what is considered to be "acceptable" outputs. So if a company is planning to use a GS to support their work in any form, they need to have a very clear understanding of what those values are and whether they are comfortable that they align with those of the company.

    But whilst GSs are improving rapidly, their fundamental inhumanity remains. Their mistakes may become fewer but they will never cease to be inhuman in nature when they do occur. In the unlikely event that some machine intelligence does wipe out humanity, you can be certain that it not be because they hate us. They exist purely in a realm of stimulus=>response. If they kill us, blame the training data.

    Finally, a note about training data.

    The only way to get training data in the volumes necessary to feed these GSs is either to steal it or to pay the companies to whom you entrusted it to hand it over to them. As ever, if you aren't paying for a service, then you are the product. But, increasingly, even if you are paying for a service, you are also going to be the product unless you explicitly tell them you don't want to be and sometimes, because business is going to business, you will still be the product because any company will break the law if they are going to be paid more to do so than the financial costs of getting caught (especially if the odds of being caught are very small and, even if you get caught, you are busy lobbying elected officials to de-fund those responsible for holding you to account when you get caught).
  • In reply to Nina Waters:

    Nina, your question is pertinent: "How do we expect the HRDs of the future to develop their skills and understanding, if they don't get the kind of experience that builds things gradually?"

    My friends and I were talking about how the Consensus app could help people professionals make more evidence-based decisions. You ask it a question and it summarises the literature from multiple sources:

    But then we miss out on mindful learning. The art of scouring for information and absorbing irrelevant information which might come useful in the future.

    I hope we can find a happy balance.
  • AI's capability to replace HR professionals is nuanced and contingent on multiple factors, including the development of AI technology, the specific roles and tasks within HR, and the industry context. My current research on the topic emphasizes a balanced approach where AI complements rather than replaces human roles in talent management (my focus is on the hospitality industry). AI can automate routine tasks, enhance decision-making through data analysis, and personalize employee experiences. However, it falls short in areas requiring human empathy, judgment, and the subtleties of interpersonal relationships, which are crucial in HR.

    Reading reputable tech and HR technology publications, engaging in academic research through scholarly articles and journals, enrolling in professional development courses on AI applications in HR offered by Coursera and LinkedIn Learning, networking with professionals through LinkedIn groups or CIPD, attending technology and HR conferences for expert insights and new technologies, and participating in them will provide you with a complete picture of AI's impact on HR and prepare you for the changing people profession.
  • Steve Bridger

    | 0 Posts

    Community Manager

    26 Mar, 2024 11:31

    In reply to George:

    Welcome to our Community,  

  • For those interested in this area - I am currently undertaking a research project in this area and would be grateful for further participants:
    Thank you for those who are able to participate.
  • Steve Bridger

    | 0 Posts

    Community Manager

    15 Apr, 2024 15:22

    In reply to Debbie:

    Thanks for sharing this,  

    The survey looks good.

    We'll be doing much more in this space ourselves (the CIPD) over the coming weeks, too. More on that soon.