ChatGPT isn't really intelligent. It has no ideas of its own – it's all based on existing ideas. It constantly "learns" by acquiring new data, including anything you post. — Image by pch.vector on Freepik
There are many amazing things about ChatGPT and other AI tools that are available. AI can do everything from writing your headlines for you to drawing you the presidents of the United States as if they were supervillains.
It's fun. But it's also useful, and you've probably been using AI already in your business, especially if you don't have a policy that prohibits employees from using ChatGPT for work purposes. And why not? If it saves time, it's worth it.
Yes, and no. While AI is great for generating headlines, you need to be prepared if you start using it to make hiring decisions.
Data privacy
ChatGPT isn't really intelligent. It has no ideas of its own – it's all based on existing ideas. It constantly "learns" by acquiring new data, including anything you post.
Craig Balding, the author of the cybersecurity and AI newsletter and blog on ThreatPrompt.com, told Worklife that what your employees submit could show up in someone else's answer: "You could have your own password showing up somewhere else. Although there's no attribution, there are identifiers like company name, well-known project names, and more."
This is a concern in all facets of business, including hiring. For instance, if a recruiter wants to give hiring managers a quick summary of candidates, ChatGPT summarizes resumes well. But now you've put someone's whole work history, address, and phone number into the data set. Did your candidates consent to this?
Bias and error
We know ChatGPT, like all artificial intelligence, is biased. It's only as good as the programmers and the datasets. Since these are flawed, the AI tools are biased. Amazon created an AI system to help with hiring and had to stop it because of bias. Illinois is suing HireVue, a video interviewing program that evaluates the facial geometry of candidates to determine "cognitive ability, personality traits, emotional intelligence, and social aptitude," because of bias.
And bias aside, ChatGPT makes things up when it doesn't know the answer. I asked it to summarize a newly released New York State case that dealt with abortion rights in the workplace, and ChatGPT returned information about gun rights.
Because I was familiar with the case before I asked for the summary, I knew it was wrong. But if you use this tool to help answer legal questions, you need an educated person asking the question. You cannot trust that ChatGPT will return the correct information.
You're legally responsible for everything your company does
It would be nice to say that any illegal bias in your hiring or employee retention efforts could be blamed on ChatGPT or another AI program, but you cannot. Whatever decision your company makes is your responsibility.
In a recent webinar, employment attorney and HR consultant Kate Bischoff discussed the important legalities you need to consider in staffing. Here are three critical things:
- The laws are always running behind the technology. Right now, things are not at all settled in the world of AI and hiring. There will be court cases. There will be new legislation. Keep your eye out.
- Do you know what your vendors are doing with AI? Do you know what happens with your data? Not only do you need a policy, but you need to know your vendor's policies.
- You are responsible for any decisions. It doesn't matter if your vendor advised you; it's still your decision, and your business is on the line. You can't push it off onto them.
You can't illegally discriminate against candidates and employees – even if it was the black box of AI that made the suggestion. You still act.
If you don't have a policy, your employees will use ChatGPT and other AI programs. Make sure you're OK with that and get your policies together. – Inc/Tribune News Service
