This article is inspired by a stimulating discussion at an HR leaders meetup at Odgers Berndtson last week and from listening to this podcast involving Stuart Russell – a pioneering authority on Artificial Intelligence.

AI and why we don’t need to worry

As with any buzzword artificial intelligence (AI henceforth) is being thrown around loosely. But let’s just remind ourselves what it is at a high level. It is any man made system that tries to take in inputs, process it and come up with some conclusions that helps in making decisions towards a goal. It works by taking lots of input data, trying to find a model that fits the data and thus use the model to make decisions in the future.

So when you ask your Siri/Alexa/Google Home a question, the AI in question within those devices is taking the input of the sound of your voice, turning it into a format it can then process (sometimes hilariously badly) and then deciding which action to take on the back of it – to put in a reminder for you, order something else you probably don’t need or do a Google search for some useless piece of information you’ll immediately forget! Sometimes it might do some combination of all three and this can be even more comedic or tragic depending on your view on life.

Thus far, most forms of AI are only good at doing one task. So when it comes to assessing someone (potential or current employee) to take on a new challenge and how to motivate them, it is unlikely to be able to replace a human.

Now this should not be news to most. But let us just emphasise this point further. Human intelligence is now widely thought to be composed of IQ and EQ (and perhaps many other components). In many industries HR practices centred around optimising for IQ and if this were the case, AI would threaten a lot of HR jobs and perhaps even beyond. However, since we aren’t just number crunchers, and there is much more to a ‘high performing employee’ contributing well to the company’s goals, AI is not as much of a threat. And our prediction is that the many dimensions across which human behaviors and motivations will be judged will only keep increasing as we get more and more data.

So what about the AI being used now

Given these quite obvious limitations of AI, it is dangerous to be using it to enable decision making on something as important as how an individual is treated with respect to their careers and work life.

But in the face of ever increasing applicant numbers, the desire to provide support and guidance across an ever increasing spectrum of responsibilities (who knew employee perks and wellbeing would take up so much time and effort!) and of course leverage data for everything, AI is the only hope. And so we have ‘shallow’ implementations – crude processes that only scratch the surface of the nuanced understanding it takes to do an HR role well.

Take a look beyond AI and the picture is all the more bleak. AI algorithms controlling social media has turned many of our best and brightest into vapid endorphin-junkies looking for the next hit from a ‘like’ from a ‘friend’ we haven’t spoken to in years. Worse, it is helping to create bubbles around us that lead to mis-information and mis-guidance becoming an ever more powerful force in our world. 

As Stuart Russell says in this podcast, it’s a case of capitalist mind sets going full throttle to exploit the opportunity created by the new technology without the oversight and regulation society often needs. And this isn’t the first time it’s happened like this – tobacco companies were able to convince most in the first half of the 20th century that smoking is actually good for them. And Exxon (apologies if you work there) did and continues to spread mis-information and doubts about climate change that really harms the prospects of the many, in favour of a few.

What can we do

As responsible, knowledgeable citizens of the world, it’s down to us all to be the voice of common sense in checking what AI technology is doing for us in the near term. Because these processes are only going to get further embedded as the algorithms get better, faster and more ubiquitous.

If you’re going to use a piece of HR tech to help make your decisions (and there are many noble and valuable reasons to do so) make sure the decisions are as un-biased as possible and not based on past prejudices. If this feels tricky, we understand – we’re hell bent on building AI powered tech that levels the playing field and we know it’s not always straightforward. Come chat to us and we promise to offer some advice – not because we have something to sell, but because this matters, and it matters now.