AI or not AI – that is NOT the question

I have written already a few posts around AI and I am sure that everyone is clear how big of a fan I am of AI – especially Generative AI. I am convinced that it is as big of a change for all things business and all things private like the internet itself – maybe even bigger. And for sure, HR is not to be taken out of this. The way we do HR, the way People Processes are delivered, the way Talent Development is executed, the way we do People Management will all change in significant ways because of AI. And so, I am not a fan of these apocalyptic future views that either will lead to high unemployment because “AI will take our jobs” or even worse “AI will destroy humanity“.

Of course, AI will replace some jobs, especially routine jobs – and it will make other jobs significantly more efficient. But like with all technologies and their business world revolutions, new jobs will be created and in the end, more value for most and more jobs for many will be the result. However, also as in all disruptive changes before, it will take time to even out and, we will leave certain groups behind. Workers that are not willing to change or that cannot change. How we deal with this is a societal question as much as a debate that needs to happen in every enterprise. 

AI or not AI is NOT the question

But, across this debate it is certainly very clear that “No AI” is not the answer – except if you are ready to be out of business soon. AI is out of the box – regardless if you call it a pandora’s box. It is here to stay (opposed to the Metaverse) and the only question is truly how you make the most use of it. And what are the right adaptions for your business. From an HR perspective – and this is the view that I am taking – there is so much to win and go after that I am not really sure where to start and where to go first. Still, before going there it has to be called out that AI is not an off-the-shelf ready to apply technology (yet). There are certain watch-outs that must be taken into consideration when implementing it – especially in the space of HR. 

Watch out for hallucination and bias

Generative AI still has its flaws and limitations – especially the hallucination topic that must be addressed or cared for when you use it. I don’t think it would be a good experience if someone is checking the PTO policy and the answer is double or even more of what is actually granted. It is important to cater for that of course, but much worse than that is bias. If you use AI or ML in any decision making process, it will do as good as it was trained. And never forget that it has been trained on decisions and data that were human made – with all our flaws and biases. The problem with AI is that it will further amplify the biases if you use it widely. It is important to be able to follow the decision making process of the AI and not having it a complete black-box. And that is still a big problem today. 

AI is nothing worth if not implemented with enablement

These were just a few flaws that are inherent to the technology itself. There is one additional aspect that needs to be taken into consideration – and where we have truly failed in the past: enablement. Of course, you can put AI into anyone’s hand today but if they don’t know how to use it, what to use it for, how to embed it in their daily workflow, you will end up like we do today where the majority of colleagues struggle with even Excel or PowerPoint as soon as it gets a bit more complex, e.g. using formulas, using a slide for more than bullet points… It would be a tremendous waste of investment  and innovation. Enablement is key – and yes, even if AI will be used as co-pilot to get the job done, it is important that whoever uses it knows how to use it best for this specific application of it. Everyone today can and probably has used ChatGPT or Dall-E or Bard or any of the other incarnations of Gen AI. But did everyone get the same results from it? – No. Triggering it with the right question, the right data points and syntax is key. It will distinguish mediocre results from great results. There is even a job market already for prompt engineering. The right capability is key to be successful – and will make the difference between adopters and winners.  

AI or not AI is NOT the question – how it is used and what it is used for is what makes the difference. If you are not yet running MVPs or test & learns with Gen AI you are close to missing this train. It is leaving the station and accelerating.