HR must take control of data and AI

As HR analytics continue to become a vital tool for those seeking to improve strategic decision-making and optimise their workforce, Alistair Dent, chief strategy officer at Profusion, advises HR leaders on how to strike the balance between productivity and privacy.

It’s safe to say that the HR field is becoming increasingly complex as new technology coincides with new working patterns and new agendas. With this it can be hard for HR leaders to keep pace and maintain a watchful eye as to how employees are feeling, performing, progressing and everything in between. Add to the equation the advent of a hybrid-working culture over the past three years it has made it even more difficult for HR teams to maintain an “ear on the ground” without the physical touchpoints afforded by an office environment.

As a result, it’s clear why recent years have seen the use of HR analytics increase. Done well, a data-led approach can have a transformative effect on the effectiveness and efficiency of all types of key HR operations. When it comes to talent acquisition, for example, it can help businesses identify the characteristics of high-performing employees to target their recruitment efforts more effectively. Equally, visibility into areas such as employee satisfaction, exit trends and staff turnover rates can be used to inform better decisions designed to improve staff retention, such as benefits, and training programmes. More so, through HR analytics it becomes easier to key track employee productivity and performance, especially for those working at home, and recognise and reward accordingly.

But while the rationale for data-led HR may appear conclusive, it does bring new challenges. Principally, these concern the risk of violating employee privacy and the inadvertent creation of discriminatory or opaque practices. Therefore, it is crucial for HR professionals to prioritise ethics in the use of HR analytics and achieve the right balance between data collection, employee privacy and transparency.

One of the most obvious areas of debate in this is just how influential data should be on decision making. Some may argue that when it comes to promotion or personal development, an algorithm would be a poor substitute for experience and intuition. Although, what is experience other than years of information collected by an individual and ‘analysed’ in a way that will let them recognise and respond well to similar situations? Such a position also ignores the fact that data science’s capability goes far beyond measuring simple productivity stats – it can take into account a huge range of sources from the structured to the unstructured – such as behaviour on a video call meeting. Then again, to get enough data points, does that not mean workers will be under an unprecedented level of monitoring? Perhaps privacy issues matter less if it’s an algorithm crunching numbers rather than a nosy boss keeping tabs?

Advocates of data science in HR also argue that it is inherently fairer. Rather than leaving decisions up to the personal whims or potential human error of an HR rep or line manager – give the power to a dispassionate machine. This argument is particularly powerful when it comes to finding ways to improve diversity and inclusion within an organisation.

Thinking about human or data based decision making in HR as a binary choice can lead to you getting the worst of both worlds. Though it was only a trial and didn’t go live, Amazon saw how an unchecked recruitment algorithm designed to end an existing bias against women and to eliminate the gender gap actually ended up perpetuating these trends. According to a Reuters report this issue arose partly because “Amazon’s computer models were trained to vet applicants by observing patterns in resumes submitted to the company over a 10-year period. Most came from men, a reflection of male dominance across the tech industry.”

The answer is to keep the human in HR. Data science is a powerful enhancement but it is not a replacement. Although it has the capacity to make decision-making fairer, more transparent and tackle systemic problems such as discrimination, it cannot work properly if it is left to its own devices. Amazon’s findings could have been avoided if there was closer monitoring and scrutiny of how its algorithm actually impacted the wider business.

Ideally, HR and data science should work hand in hand, with both parties collaborating in the collection of data, how algorithms are designed, how they are refined, and, critically, how the outputs are interpreted to make decisions.

For this to work, HR representatives and data scientists need to share their skill sets. An HR professional must upskill and educate themselves on the basics of data analysis to be able to properly understand its benefits and limitations, and to be empowered to scrutinise how it functions in a business. Data scientists will be better able to design and build new HR solutions if they fully understand the needs of HR and a company’s employees. Alongside this, a system of checks and balances can be developed – by both parties – that maximise its benefits and minimise its risks.

The final piece of the puzzle is communicating with the wider business. Data-driven HR is only fairer if it is a transparent process that everyone understands. This may require company-wide education on data to demystify the whole process. The benefit is that it will empower people to challenge decisions they feel are unfair – adding further safety and accountability on the system.

Alongside this, as HR continues to double down on its approach to data ethics, the hope is that legislation does too, and at pace. A new data bill is now at Report Stage in the House of Commons and looks likely to be passed ahead of the next General Election. Thus, as new technologies such as ChatGPT bring fresh possibility and concern, and despite new laws, the onus looks likely to remain on businesses to go beyond compliance and do the right thing. In this way, through planning, transparency, employee engagement and consistent human intervention, HR leaders can harness the huge productivity gains to be had by HR analytics while keeping their people on side.

Latest HR job opportunities on Personnel Today

Browse more human resources jobs