You should never let an intern use ChatGPT.
Or any AI for that matter. Take a look at this scenario ChatGPT created:
Maternity Leave Misclassification
Scenario: A pregnant employee starts having complications at 24 weeks and goes on FMLA. HR incorrectly starts counting the 12 weeks from this date, leaving no job protection for post-birth recovery.
How HR Might Get It Wrong:
Misapplying the 12-week entitlement by not recognizing that pregnancy complications and postnatal recovery can trigger separate blocks of qualifying leave.
The problem: This makes it sound like the employee is entitled to 12 weeks of FMLA for pregnancy complications and then an additional 12 weeks for birth and bonding.
Now, are there times this would be true? Sure. If you operate on a fixed calendar year and the year switches over smack in the middle of this, she could get 24 weeks of protected leave (assuming she worked 1250 hours prior to the second leave).
Could PWFA require you to provide her additional protected time off even if she used all 12 weeks of FMLA during the pregnancy? Absolutely.
But the AI answer is misleading and suggests that leave during pregnancy is not covered by FMLA. It makes a complicated situation seem simple.
I love AI. But you can’t trust it as an expert.
I’m somewhat of an AI expert in my field. You should NEVER use AI when you don’t know enough to be able to spot and correct errors.
An AI can speed up an experienced employee. But if you *need* the AI, you shouldn’t use one.