Posted on

Generating job descriptions

In the previous module, we saw some examples of how a general LLM can be used to assist in generating job descriptions.

There are also platforms tailored to the process. For example, AI-powered tools fromTextio can analyse the wording of a job post and suggest edits to make it more inclusive and appealing to a broader range of candidates, thereby improving the diversity of applicants.

Textio is a tech company that started in 2014. They highlight how often unconscious biases can affect inclusivity, particularly in job postings. Language that seems neutral on the surface, such as “exhaustive” and “fearless,” can discourage female applicants due to subtle gender biases. Similarly, words and phrases rooted in ageism or ableism, along with expressions that differ culturally across regions, like “bi-weekly,” and business jargon stemming from historically male-dominated corporate environments, like “stakeholders” and “KPIs,” may unintentionally alienate potential candidates.

Moreover, phrases with racist origins, such as “grandfathered in” or “blacklist,” continue to be used despite their problematic histories and the presence of more neutral alternatives. The use of language that generalises or labels individuals, like referencing someone as a “diverse person,” can also have negative connotations.

They argue that fostering a truly inclusive workplace, the focus should not just be on eliminating exclusive language but on adopting language that actively expresses a commitment to inclusivity. This can be done explicitly by stating equal opportunity stances or implicitly through phrases that convey inclusive values.

Creating such a culture involves appealing to individuals who appreciate diversity, are community-oriented, and view talent as something that is developed through experience and effort. Focusing on these attributes can make underrepresented groups feel more included and connected, building a self-reinforcing loop that sustains inclusivity within the organisation. Achieving this requires an ongoing commitment to consider the impact of language on talent attraction and inclusivity efforts.

They offer a range of tools that use AI to generate documents like job descriptions that aim to improve the inclusiveness of language as well as provide efficiency benefits. They provide three categories of language guidance:

  • Multicultural: The tools highlight phrases that reflect a multicultural approach to promoting diversity such as “We embrace diversity and want you to bring your whole self to work.”
  • Team-centric vs. individual achievement: Phrases that emphasise the importance of teamwork and community are prioritised.
  • Growth vs. fixed mindset: The tools aim to move the language away from fixed mindset phrases and towards language that emphasises learning and growth. Instead of saying your team is “super smart,” use phrases like “dedicated” and “resourceful.”

The tools use colour highlighting and drop-down suggestions to guide the user through making a stronger written piece. Although these tools can improve inclusion and leave ultimate choices to the writer, they will still tend to result in documents that present a Textio view of the world, which may not align perfectly with your organisation. It may also reduce the individuality and personality of documents read by potential recruits.

Would you use AI to help you generate job descriptions? What benefits and challenges do you envisage?

Record your thoughts in your learning diary before you move on….