At the moment, Generative AI, while interesting, novel, and perhaps a wild card, is not the digital John Wick that it is often made out to be, coming to end the jobs of humans everywhere. Conversely there are headlines that tout that AI is without bias.
The provocative polarization by these sort headlines meant to get clicks serves only to reinforce the echo chambers that seem to be the norm today. As HR professionals, we need to have a more nuanced understanding.
In viewing this through the lens of a human resources person who focuses on recruiting for small start ups, the real question is whether tools like ChatGPT, Gemini, Claude et al can truly be a force multiplier and allow me to do more in less time.
Based on the current state of AI, the simplest, least glamorous answer is … it depends.
At the moment, the output of chatbots is based on the strength of the prompt you create. That said, writing prompts can be a little like making a wish and rubbing Aladdin's Lamp. If your prompt is not well built, you will get a result that is likely not useful.
Imagine then, you are a solo HR practitioner in a startup needing to craft a paragraph on the Pay Transparency Act and you want to see what other companies have done. Could you create a prompt that allows you to mine for that data? You bet.
As with any skill, writing an effective prompt takes time to master. So is it faster to compile the information by just grabbing three or four job postings, scanning for the data you need and writing the paragraph? Probably.
So does this mean game over? No need to learn ChatGPT. Nope. Not even close.
With any new technology the initial investment time is high to learn to use it effectively. And with constantly conflicting priorities, it can become easy enough to put AI at the end of the list. This is where as an HR professional you need to take a different view.
HR is uniquely positioned within a company to take the lead and partner with other groups to implement a way to utilize AI for the benefit of the business while being socially responsible.
In order to do that, as HR people, we need to:
Reframe the Conversation: AI may present as a technical issue, however it is actually a human issue. It impacts the workforce and presents both opportunities and challenges.
Keep Current: Large Language Models like ChatGPT are still in their infancy and will only get better. Pay attention to those improvements. However be aware of what's coming next such as Large Action Models (LAMs). LLMs, like the ones we've discussed, just generate text that you still have to act on. LAMs, on the other hand, focus on taking action in the real world. So, an LLM might help you write a compelling job description, but an LAM could potentially automate the process of posting that description across multiple job boards. Of course, there might be some initial work involved in setting up these connections, like integrating the LAM with your company's applicant tracking system. But the potential time-saving benefits in the long run are significant.
Become AI Literate: invest the time to understand how to write effective prompts. Understanding this aspect and becoming proficient at it allows us to assess what value it can bring to our own practices. Not only do we need to become literate, we also need to encourage and create opportunities for employees to do so as well.
At this moment, Generative AI is still in the hands of the early adopters. Innovation is still taking place to determine effective ways to use the current tools. Many of us have already experimented with some version of a prompt “make this sound better” or “write this performance review.” Fewer of us have engaged in more complex prompts like harvesting the Pay Transparency Act language from different job postings.
Candidly, now is the time to play around. Creating a good prompt is less about knowing how to program, and more about how to logically write a set of instructions. So experiment. Try different tools.
Every tool has its merits. For example Google’s Gemini, allows you to craft a compelling role-play between yourself and a digital employee. The value of being able to practice a difficult conversation can be very helpful preparation.
For example: “I need to counsel an employee whose co-workers have complained about his attitude. Can you roleplay the employee in this scenario?”
However in order to be truly consultative, you must also be aware of current limitations. AI does have bias. Full stop. Period. Factoring that knowledge in can allow for more informed choices about how you might deploy an Applicant Tracking System.
And for now, remember AI is just good at correlation, not causation. For example, how many of you have ever received an email for a role that was not related at all to your experience? So the search agent that was created, caught the keywords, but not the context in how they were used. "Yes. I have Java on my resume because I worked at a coffee shop, not that I know how to code."
Our intelligence to create these tools typically exceeds our wisdom in using them.
Human Resources professionals just like you can ensure that there is nuanced and thoughtful conversation around both the benefits and detriments of using AI in your business. Being consultative and sparking conversations with senior leaders and employees alike will demonstrate the business value that HR professionals can bring to any company.