It is now almost a year since OpenAI surprised the world with the launch of their ChatGPT AI chatbot. Since then, Microsoft, Google and others have launches a range of AI tools with capabilities comparable to ChapGPT. Professor Andy Charlwood discusses his early findings on the impact of AI on HR since then.
In April 2023, the ThinkTank launched a survey to find out how these tools were being used by HR and HR analytics professionals. Our survey results suggested that few people working in HR and analytics had adopted AI tools at that point. The CIPD recently invited me to lead a discussion on the impact of AI on HR work at their annual conference and it was immediately clear a lot has changed since April.
HR's new co-pilot?
I was blown away by the number of HR professionals telling me that they are using AI tools in their day to day work. AI tools are being used in a “co-pilot” role to augment capabilities and enhance productivity. One HR manager in an SME told me they were using AI to draft emails and routine employee communications, resulting in significant improvements in quality and productivity. An HR analyst told me using Power Automate had dramatically increased her productivity as she was able to automate routine data preparation and analysis work, while GPT-4 was great for speeding up coding tasks. Users of AI tools reported that the extra time AI was giving them was allowing them to focus on more strategic and impactful activity within their organisations.
QuoteI was blown away by the number of HR professionals telling me that they are using AI tools in their day to day work.
These anecdotal experiences chime with emerging academic evidence on the impact of AI on quality and productivity. One recent Harvard study has found that management consultants who use AI are able to complete 12% more tasks because they are executing those tasks 25% quicker. Below average performers benefited most from using AI. All of those using AI were performing above the pre-AI performance average by the end of the trial.
Care is needed
However, care is needed when using AI. The HR professionals I talked to raised questions about which tasks could be delegated to AI and which still required the human touch and human judgement. They also pointed out that when working with HR data, it is important to think about data protection and security. We can take things (code, text) out of AI tools, but we need to be careful when framing queries and prompts not to add in any of our organisations confidential data. It can also be hard figuring out which tasks AI is likely to be good at. As Wharton Professor Ethan Mollick wrote recently, AI is “weird”. No one really knows the full capabilities of tools like GPT-4, and there is no instruction manual to consult. AI can easily complete some complex tasks, while failing subtly or spectacularly at others. Experience is needed to understand what it can be safely used for.
GPT AI tools are also throwing up new challenges for HR. One recruiter told me that CVs and covering letters are less effective in identifying candidates to bring forward to interview because candidates are using AI to come up with more impressive application material than they would be able to produce by themselves. Another said that they had abandoned online interviews because they were concerned that candidates were using GPT to come up with answers to questions.
ThinkTank research continues
I am currently working with Maarten Renkema and the ESRC Digit Research Centre to understand more about how AI is changing the nature of HR work. I can’t wait to share our findings with the ThinkTank in 2024 as we find out more.
