By Peter Weddle, Founder & CEO TAtech
Despite their importance to online advertising effectiveness, job ads have long been … well, mediocre or worse. All too often, they’re poorly written and about as engaging as a brick. I suppose that’s why so many pundits and evangelizers are now giddy at the prospect of AI-written ads. There’s no doubt that large language models such as ChatGPT can improve what’s now produced and posted online. They do not, however, take the human out-of-the-loop. Instead, as so often happens with new technology, they move humans into a new and very different role. Whether they’re a recruiter on a corporate recruiting team, a customer service person working for a job board, or a copy writer at a recruitment advertising agency, this new assignment is to take the ad from good to great.
The history of online job ads is a tortured one. I began publishing WEDDLE’s Guide to Employment Sites on the Internet back in 1999 (yes, Virginia, there were job boards in the 20th century), so I’ve read a lot of those ads. In the beginning, they were mostly print ads repurposed online. The real estate on newspaper pages was expensive, so recruiters made due with the briefest of information about the job written in the tersest possible language and with lots of acronyms. That’s the earliest postings on the web looked like, as well.
As the years went on, ads did get a bit better, but not much. They provided more information, but were not more informative. In far too many cases, they weren’t ads at all, but simply position descriptions, a format developed for job analysis and compensation ranking, not recruiting. That format also brought the introduction of words only a lawyer could love: Requirements and Responsibilities. Once again, the ads had little or no ability to engage anyone but the most determined job seeker.
So now, we’re in the ChatAd era. As I’ve written before, the jury is still out on whether these large language models (LLM) can actually produce good ad copy. How can I make such a statement? Because many of these models are trained with historical data. If job ads have historically been mediocre, then the models will learn how to write exactly that kind of ad.
But, there is the possibility of change. The best of these models are also trained with current usage data. In other words, if they bring back ad copy that a human determines is mediocre AND those humans then teach the model to rise above that standard, the models will perform better on the next ad and the next ad after that. In other words, LLM-produced ads can improve and I believe already are, which means that in very short order, they will produce ads that are the equal of anything recruiters can write.
Early Bird registrants get deep savings and a spot at LIMITED SEATING TAtech North America & The World Job Board Forum. Avoid FOMO, register right now for the only conference in the North American market that is specifically designed both for job board and talent technology company CEOs, their direct reports and rising stars and to accelerate the bottom-line growth of their enterprises. Why the rush? Because the conference will be held June 4-6, 2024 at The International Spy Museum which has a limited seating capacity and Early Bird Discounts on registration are due to expire shortly. So, register today and accelerate the success your company achieves, today, tomorrow and into the future!
This impending development means that recruiters must now perform a new role. The job ads written by ChatGPT and its LLM siblings will be good … but not good enough to attract the best talent. Only great job ads can do that. They’ll have the right content in the right format, but no sizzle. So, recruiters must now take on the job of editor to raise them to that higher standard. With apologies to Jim Collins, they must take job ads from good to great.
What exactly does that mean?
There are any number of definitions online for the role of editor. I like CareerExplorer’s best: “An editor's primary goal is to ensure that the content is accurate, clear, and well-organized, and that it meets the intended audience's needs.” It is the rare definition that acknowledges an editor actually has two tasks. One is to review the content to make sure it’s factually correct and easy to read and understand. The other is to make sure the content connects with and motivates the reader. Or to put it in the vernacular, to make sure it sizzles.
Even the smartest machine can’t do that. Until we reach AGI or artificial general intelligence (which as I note in my book 2118: What Humans Will Do When Machines Take Over is still a ways off), even the best language model will be unable to empathize with and relate to humans. So, that will be the primary role of recruiter-editors – to take what these smart machines produce and transform it into copy capable of convincing candidates to do the one thing we humans most hate to do: change. To go from the devil they know – their current boss, employer and commute – to the devil they don’t know – a new job.
Only great job ads have that kind of sizzle, and only recruiter-editors can create them.
Food for Thought,
Peter Weddle has authored or edited over two dozen books and been a columnist for The Wall Street Journal. He is the founder and CEO of TAtech: The Association for Talent Acquisition Solutions.