Employer Branding in the Age of AI

Everyone, it would seem, is talking about AI and the effect it will have on employment, our culture, and our lives. ChatGPT was released in late 2022 and quickly entered the public consciousness. Since then, New York Times journalists have had uncomfortable conversations with the Bing chatbot, a Cleveland-area radio station started experimenting with AI radio hosts, and image-generating sites like Midjourney and DALL-E became known for both eye-popping visuals and important questions about whether machine learning constitutes copyright infringement. Bill Gates himself has declared AI to be “the most important advance in technology since the graphical user interface.” Reading these stories can give the impression that the world changed a few months ago—and maybe it did.

When there are sudden technological developments like this, it can be hard to tell what we need to do and when we need to do it. Expectations of these new technologies are very high—and in some respects are probably overblown (at least for now). Anyone who sets out to educate themselves on this topic will encounter a mixture of fact, misinformation, and wild speculation, so what does an employer brand manager need to know about AI? 

The Basics: What are these AI tools? 

AI technology has been in development for years, but it only recently was made available to the public. Unless you were an AI researcher, you probably had no idea that the state of the art was so good—which partly explains why the release of ChatGPT created such a sensation.

ChatGPT, the Bing chatbot, Midjourney/DALL-E, and related tools all fall in the category of “generative AI,” which refers to AI that can create something on its own—an article, a song, a painting, or nearly anything else. GPT-3 and GPT-4 are LLMs (“large language models”) that are trained on an enormous amount of data; reportedly GPT-3 was trained on 40 terabytes of text data, or the equivalent of 100 billion pages of text from websites, books, blogs, etc. ChatGPT was then built on top of GPT-3 by adding a human element: first GPT-3 was fed thousands of prompts and asked to give several responses to each one, and then volunteers ranked every response by quality. Those rankings were fed back into the model to help it learn how to give the most human-sounding response.

The method that LLMs use to create a response to a prompt gives a fascinating view into what can make these tools both magical and somewhat unreliable. They all rely on statistical tables, created when they crawled those 100 billion pages of text, that describe how words and phrases are related to one another. When you enter a prompt into ChatGPT, it looks at the collection of words you entered, compares it to the statistical table, and constructs its reply one word at a time. First it picks the best initial word, and then the best word to follow that word, and so on, as if it were threading beads onto a necklace. 

This method of building a response explains why ChatGPT, Microsoft’s Bing, and Google’s Bard are all known to make factual errors from time to time. These models can only tell you what seems like the most probable answer given the prevalence of words and phrases that appear in the sources they were trained on. In other words, they don’t actually know any of the answers; they are simply making educated guesses, and sometimes those guesses are wrong.

The Recruiting World Is AI-Curious

Nevertheless, the world of recruiting and hiring is already seeing the impact of these new tools. 75% of recruiters and hiring managers hope that generative AI systems like ChatGPT can take on repetitive tasks and free up humans to do more strategic work, and two-thirds are counting on AI systems to identify strong candidates quicker and more effectively than a human can. Along the same lines, LinkedIn has announced new AI-assisted tools for writing job posts, collaborative articles, and personal profiles.

Some businesses are forging ahead with generative AI not only as a solution for their business problems, but as a replacement for some of their staff. 36% of Americans are concerned that technological advancements like AI, automation, and robots will negatively affect their career opportunities, and those concerns are even higher among younger and minority workers. They have good reason for feeling that way: 25% of employers state that they have already eliminated jobs because of ChatGPT. A recent study estimated that 80% of the US workforce will see their jobs affected by GPT, with higher-wage jobs affected to a greater extent than other segments of the workforce.

There is no question that the technical roles at your company will be affected. Machine learning will be a primary target for recruitment, and the new tools may expand the candidate pool by lowering the barriers of entry to technical roles with enhanced training and reskilling options alongside AI-assisted code development. Almost certainly, prompt engineering—the art and science of creating an effective prompt to feed into generative AI—will be an important job skill for years to come.

So, What Should You Do Now?

As employer brand professionals, we are all in the same boat—almost certainly these technologies will affect our jobs. You might soon be using generative AI as a resource when you’re writing blogs or social media posts. You or your partners in HR might use AI tools to generate effective job posts more quickly. There are already several services promising these capabilities.

With all these platforms promising they can automate your writing and marketing, should you be worried that the machines might replace you entirely? Probably not yet—the technology still has limitations. The best model is still a partnership; the AI can produce a quick first draft or apply a certain style to something you’ve already written, but you’ll want a human being to inspect the copy before anything goes live. Brands that get ahead of themselves with generative AI and one-click, no-look marketing solutions will sooner or later find themselves apologizing for errors—factually-incorrect content, inadvertent plagiarism, or outright copyright infringement.

That being said, this is a great time to roll up your sleeves and get comfortable working with generative AI (you can start with the free AI courses now available on LinkedIn). This is also a good time to get your team together and reflect on some important questions about your company’s employer brand within this new technological landscape:

  • How will these technologies affect your open roles?
  • How do they affect the candidates you will pursue?
  • How will your company support employees who might regularly need to reskill as new technology impacts their roles?

Some employers already appear to be factoring AI into their plans for workforce reduction. That message is rippling across the job market today, and it presents a golden opportunity for other brands to lead with the message that technology is best used when it enhances and extends our work, unlocking new superpowers inside our companies and making us capable of so much more.

Relationships are tested at moments like this. Candidates and employees will know when they’re dealing with a brand that thinks of them as a number and considers them to be replaceable. You can, though, leverage the same technologies to craft better, more compelling messages that connect powerfully with your audiences and strengthen their relationship with your brand.

Many years ago, Steve Jobs described the personal computer as a bicycle of the mind—technology that expands the range of human possibility and allows us to think bigger and do more. Generative AI is that sort of technology; its capabilities today are eye-opening, and tomorrow they could be transformative. This is a fast-moving landscape, and an important opportunity to educate ourselves on what these new tools are good for, assess where their weaknesses lie, and get started building a future of work that brings out the best in everyone. 

We have a brand-new bicycle of the mind. Let’s see how far we might go and where the limits will be tested.

Related Posts