?ChatGPT, a generative artificial intelligence (AI) tool, can help employers with myriad HR functions, including writing job descriptions. Indeed, I have tested the tool and found it quite helpful as a starting point for job descriptions.
However, ChatGPT doesn’t replace or otherwise remove the need for HR thinking. There are many issues AI does not address, including those discussed below. When crafting job descriptions, HR should consider the following issues using authentic, rather than artificial, intelligence.
Minimum Job Requirements
To increase the diversity of the applicant pool, we need to consider what existing minimum requirements are not actually needed and may narrow the number of applicants unnecessarily. For example, more employers are putting the college degree on the chopping block.
For some positions, a college degree truly may not be necessary. But for other positions, it is job-related and further serves as an objective basis to determine who should or shouldn’t get an interview. Remove the objective requirement and it is much easier for an applicant who wasn’t interviewed to bring a viable claim of unlawful discrimination.
If we focus only on diversity, what we may increase is litigation. So be sure to balance increasing diversity with mitigating legal risk. AI cannot find the balance for you.
Essential Functions
If there is an accommodation request, an employer may need to modify how essential functions are performed, but it does not need to eliminate them.
By contrast, if a function is nonessential, as part of the interactive dialogue to identify an accommodation, an employer may need to eliminate it.
Divide functions between essential and nonessential carefully, or else list only essential functions. AI cannot determine what is essential to your positions.
Physical Requirements
For Americans with Disabilities Act purposes, we know we should list what physical requirements are essential to a job.
For blue-collar jobs, employers generally do this well. However, for white-collar jobs, job descriptions often fall short.
In the latter scenario, consider what I would refer to as stamina requirements. For example, the employee must be able to work long days or travel regularly. Of course, you need to spell out what is meant by “long” in terms of days or “regular” in terms of travel.
These stamina requirements should be in the job description. They are not commonplace in AI.
Behavioral Competencies
If screening is done properly, most white-collar employees don’t fail because they lack the necessary technical knowledge or skill. Instead, their failure is more likely due to a deficit in a behavioral competency.
Applicable behavioral competencies should be included in the job descriptions, and behavioral questions regarding them should be asked in the interview process.
I have developed a list of more than 100 behavioral competencies. Here are just some that begin with the letter A: accessible, ability to shift gears quickly, accommodating, accurate, action-oriented, adaptable, agility in thinking, amiable, analytical and assertive.
Think of employees who have excelled or failed. What behavioral competencies did they crush? Which ones crushed them?
Your answers to these questions will help you develop behavioral competencies that should be considered not only in the hiring process but also the evaluation process. Again, they are not commonplace in AI.
FLSA Exemptions
While a job description doesn’t determine whether an employee is exempt according to the Fair Labor Standards Act (FLSA) or state law, it is a starting point in any investigation. In this situation, what is often relevant is what is not in the job description.
Let’s assume you want to argue that a particular manager is exempt under the FLSA’s executive exemption. What’s the message if the person’s job description is silent on people management requirements? Silence speaks volumes, so the job description should include such things as managing performance and completing evaluations.
Similarly, it is important that the job descriptions of employees who will be classified under the FLSA’s administrative exemption include specific responsibilities that involve discretion and independent judgment with respect to matters of significance. Of course, the requisite independence and significance must exist not only on paper but also in reality.
Finally, it is generally best to avoid assigning percentages to various tasks of an employee’s job. I try to avoid the word “always,” so I will say that percentages are almost always wrong. Including them does not guarantee an exemption but does guarantee a credibility issue.
However, there is a possible exception in California. While the primary duty test under the FLSA does not include a percentage requirement, under California law, more than 50 percent of the exempt employee’s time generally must be spent on exempt duties.
Consequently, for California jobs, employers may wish to include a general statement of what tasks combined equal more than 50 percent of an employee’s time but not assign specific numbers to each individual task.
Reliance on AI for wage and hour compliance will serve only to make plaintiffs’ lawyers rich. Specific facts matter!
Pay Equity
In many states, two jobs do not need to be the same for pay equity purposes. Instead, they must be comparable, substantially similar or the like. These states include California, Colorado, Illinois, Maryland, Massachusetts, New Jersey, New York, Oregon and Washington.
Job descriptions often are the starting point for determining comparability or similarity. Here’s where software used to populate job descriptions can be problematic.
Many employers use software to help ensure consistent terminology in job descriptions. Such programs also help save time.
But the software can do one more thing: It can increase the likelihood that jobs will be deemed comparable or substantially similar for pay equity purposes.
You don’t want your job descriptions to force positions into a group of comparable or similar positions. Therefore, you want to start with what positions you believe in good faith are comparable or similar for pay equity purposes and then use the software to import comparable job description language for them. The concerns regarding software apply equally to AI.
The bottom line is that how employees are grouped can help or hurt a pay equity analysis. Use your job descriptions to help differentiate your groups, as opposed to having the descriptions force jobs into groups.
Conclusion
None of the issues addressed in this article can be found on an AI tool specific to your workplace. Again, this does not mean AI cannot be a helpful tool in crafting job descriptions. It most certainly can be. But as with other uses of AI, it is a tool and not a substitute for HR intelligence.
Jonathan A. Segal is a partner at Duane Morris in Philadelphia and a SHRM columnist. Follow him on Twitter @Jonathan_HR_Law.