TECHNOLOGY
A nalysing A I A pplications
Mathew Akrigg ACIPP, CIPP policy and research officer, explains how generative artificial intelligence (AI) can be used in a beneficial way, while reminding us to be vigilant to its limitations
S ome AI articles are all doom and gloom about how it’s going to take your job and make you redundant. Not this one. There may be a time where payroll is completely largely by AI, but I feel that time is some way into the future. What we need to do now is to get over our fear and distrust of it and start to harness it in small but meaningful ways. “Robotic process automation is not artificial intelligence” An introductory note Before we really dive into it, I’m going to get something off my chest, something that seems pervasive in the media and payroll industry when it comes to AI. Robotic process automation (RPA) is not AI. To explain what I mean, imagine you have an actual robot working in a car factory. It moves pieces of metal into place and welds them with precision. It’s smart, but it isn’t intelligent – this robot is exhibiting RPA. It does one job and does it well, but if there’s no piece of metal to pick up and weld, it ceases to be useful. To make the jump to intelligence, in this case, the robot must look at context and surrounding information. Most payroll AI I have seen in the market actually falls into RPA in my opinion.
To give an example of AI in payroll, I saw a demo where the system flagged something to the user. It said something to the effect of: ‘you normally send this payroll for approval by 4pm on Thursdays, are you still awaiting further timesheets?’ In my eyes, this goes beyond automation, it’s not a simple ‘if this, then that’ piece of programming. It is instead looking at contextual information and historic data, to provide a recommendation or prompt that normal programming otherwise wouldn’t. Enough of my rant. This shows the direction of AI in payroll as being led by software and solutions developers. But what can you do yourself to embrace AI and implement it into your workflow? Generative AI Well, the obvious candidate is generative text-based AI, such as ChatGPT, Google’s Gemini and Microsoft’s Copilot. These are the types of AI you can talk to, and they reply in an eerily human way. Many people are using these as tools to better their understanding of a topic without having to trawl through pages of search engine results. The AI has already done that for you! All you need to do now is ask the right questions. Such as: ‘hi, can you really simply explain to me the UK statutory sick pay eligibility criteria? I have looked at the government’s guidance and I’m confused.’ Here’s what ChatGPT had to say to that: “Statutory sick pay (SSP) in the UK is a payment made by employers to
| Professional in Payroll, Pensions and Reward | April 2024 | Issue 99 50
Made with FlippingBook - Online magazine maker