0:00
HR Party of One is brought to you by BerniePortal.
0:03
Remember the bygone era of 2022, when everything we read was nearly
0:07
guaranteed to be written by a real person? Now, we wonder if news articles, books, music,
0:14
emails, and everything we interact with daily are penned by human hand or machine. The truth is,
0:21
it’s hard to tell these days. And it’s only going to get harder. Generative AI
0:27
scrapes every corner of the internet to produce detailed, contextual results. It
0:33
can produce incredible work in no time at all, and most programs are free to use.
0:38
Imagine if we had all had something like ChatGPT in school. I can think
0:42
of quite a few papers I could have used AI to write for me! But would high school-aged
0:48
me have done my due diligence in using AI properly, maintaining ethical standards,
0:54
and fact-checking thoroughly? Or would I have just reviewed the paper quickly,
0:58
submitted it, and moved on with my day? What would your employees do… ?
1:02
As a human resources professional, you know when something might be too good to be true.
1:08
So, in this episode, let’s look deeper, so you can make the best
1:12
decisions for your organization. We’ll cover: The Pros and Cons of Using Generative AI, and,
1:18
3 Defining Questions to Ask Before Writing a Generative AI Policy.
1:23
Let’s get started!
Cons of Generative AI in the Workplace
1:24
The Pros and Cons of Generative AI in the Workplace.
1:28
AI presents an obvious opportunity in your business—but is it safe to use?
1:33
As an employer, you’re thinking about long-term success and viability. So,
1:38
let’s cover the most impactful cons, and then I’ll share the
1:41
pros that keep your organization’s future growth and stability in mind.
1:46
Using generative AI invites increased legal compliance risks.
1:51
You may have seen our episode on using AI during recruitment, so you know some of
1:51
the risks inherent to artificial intelligence in your own role. However, AI is advancing rapidly,
1:53
and regulations require updates as soon as they hit your radar. How someone uses
1:58
it today may be illegal in a week, so companies must remain vigilant.
2:03
It can produce content that may be biased, offensive, or otherwise inappropriate.
2:08
If people use it carelessly, your organization’s reputation could take a hit when AI-generated
2:13
tools say something bad. This does not just affect companies that lean into AI to produce
2:18
blogs or content. It’s easier these days for organizations of any size to deploy chatbots
2:24
to reduce the strain on service teams. If your organization uses one to interface
2:30
with clients or customers, ensure it is carefully monitored.
2:34
In 2024, delivery firm DPD’s chatbot insulted and swore at a customer seeking help. Then,
2:42
the AI-powered bot produced poetry on how bad the company is. Pretty crazy,
2:47
but I guess that’s how you lose customers in the 21st century—villainous chatbots.
2:53
Protecting sensitive data is more challenging.
2:56
Employees using generative AI tools might inadvertently share sensitive information.
3:02
Not much is yet known concerning generative AI and cybersecurity efforts. Imagine how a
3:09
HIPAA compliance officer would react if a nurse recklessly installed and used AI to
3:15
summarize patient charts, exposing hundreds of people’s private medical information.
3:21
Questions of ownership may arise. AI-generated content can raise questions
3:26
about ownership, particularly surrounding complex intellectual property (IP) laws.
3:31
AIs train themselves by pulling data from every corner of the internet. AI business
3:35
experts Scott Snyder, adjunct at Penn Engineering, and Mark Pecen, Wharton alumnus, consider AI and
3:37
patents the new battleground for business leaders. If AI provides code to a small software company,
3:44
who’s to say it does not accidentally infringe on patented technology from a much larger company?
3:49
While such a case may be ultimately unfounded, a small business cannot compete with multinational
3:55
firms in court—they’ll just wear you out with legal fees. This scenario, and many
3:59
others, provide ample opportunity for a potentially fatal misstep. Tread carefully!
4:06
Ensuring quality and consistency may be trickier. AI-generated content might not always meet your
4:12
company's quality standards. Look, if you have a marketing team, they are probably
4:17
on top of this already. But if your org is smaller, think about how much inconsistent
4:22
content and branding can hurt in the long run before using generative AI left and right.
4:27
Training is likely going to be necessary. Prompting AI isn’t difficult, but there
4:32
are tricks to it that users should learn ahead of time. Additionally,
4:36
think about how AI isn’t always correct. You’ll need to be absolutely positive people using AI
4:43
can also comprehend the content and confirm its accuracy. I like to reference the time when one
4:48
lawyer used ChatGPT to prepare legal filings for a client, and the AI invented realistic-sounding
4:55
court cases to establish precedents supporting the case. The judge was not happy, to say the least,
5:00
and neither was the client! While all of this is compelling
Pros of Generative AI in the Workplace
5:04
evidence against using generative AI EVER, much less where you work, the pros can’t be ignored.
5:11
Generative AI can optimize efficiency and productivity.
5:15
You just can’t get around the fact that it takes AI only seconds to do work that
5:20
takes a real person hours. Increased time savings enable workers to focus on more
5:26
high-value projects. It can also automate processes prone to error, reducing risks
5:31
and streamlining workflows. And, for your people who are just plain awful at writing,
5:37
AI is an obvious solution. No more client complaints about poor email etiquette!
5:42
It can result in organizational cost-savings. AI can automate expensive workflows in your
5:48
business. For example, you can use AI to evaluate benefits plan options before renewal, comparing
5:54
brokerage agencies and costs to help you save. Some AI programs may provide a cost-effective
6:01
solution if you need to hire a translator to format documentation in another language.
6:06
AI doesn’t require a lengthy interview process. If you’re struggling to hire for certain roles,
6:12
generative AI may present a solution to at least solve the issue temporarily. A small law
6:18
firm struggling to hire and retain a receptionist may be able to lean on a program using generative
6:23
AI to direct phone calls, take notes, etc. It could sharpen your business’s competitive edge.
6:29
If your competitors use AI—and I promise they do—then their output,
6:34
efficiency, and budget are all on the rise. But yours could be, too. For SMB,
6:40
especially, AI levels the playing field. It sparks innovation and enhances efficiency,
6:45
enabling smaller companies to gain entry to new markets or broaden their audiences much
6:50
more easily. On a sneakier note, you could use this uncertainty in the labor market
6:56
to scoop the excellent talent getting let go prematurely due to the sudden infatuation with AI.
7:03
So, the pros are indeed more compelling than the cons. And let’s be realistic here for a second:
7:08
your employees already use generative AI. Have you noticed lately that anyone is
7:14
producing higher quality work in a fraction of the time…? Writing a policy to ensure the
7:19
proper use of AI isn’t just a future project; it’s one you should think about doing NOW.
3 Questions to Ask Before Writing a Generative AI Policy
7:26
-3 Defining Questions to Ask Before Writing a Generative AI Policy.
7:30
When you’re creating this policy, it’s important to be vague. Being vague has
7:35
its strengths in your line of work—it allows you to decide things on a case-by-case basis. Plus,
7:41
with how quickly AI is advancing, it’s honestly just better not to get too
7:45
specific. You’ll probably need to reevaluate this policy more often than others. So,
7:50
let’s instead set parameters around AI use to create personal accountability.
7:56
Here are some questions to help you get started:
7:59
Are there times when you would consider using AI to be inappropriate?
8:03
For example, it makes me feel uncomfortable to read a news article about a child losing
8:08
their life that a real, empathetic person didn’t write. If your org handles private information,
8:14
then it’s not only inappropriate to install an unsafe AI helper; it’s illegal and risks
8:20
your business. Consider how people can use AI in your organization and set firm boundaries.
8:26
Is there such a thing as relying too much on AI?
8:30
People must exercise their best judgment when using AI. My teams know using it to
8:35
brainstorm and help with tricky projects is fine, but anything more is disingenuous to our
8:42
mission to provide the best HR resources and first-class, personalized service.
8:47
Is AI-produced work being thoroughly fact-checked? Since AI-powered writing tools hit the web,
8:53
people have been using them to quickly and cheaply write books to sell online. In one such case,
8:58
someone was using it to write guides for people learning how to forage wild mushrooms
9:03
in the forest. However, the book was wrong. It mislabeled fatal fungi as perfectly safe
9:09
for consumption. The seller wasn’t a biologist, did not have the expertise to identify mistakes,
9:16
and the dangerous claim was published online. Don’t you think any half-competent
Final Thoughts
9:21
lawyer could claim the e-book seller was responsible for manslaughter, if not worse,
9:26
if someone followed the fatal guide’s words?
9:29
As you create this policy, remember to ensure everyone reads it and signs their agreement to
9:34
follow it. Collect and store those signatures safely—they prove your organization has done
9:39
its due diligence to prevent employees from misusing AI. With BerniePortal,
9:44
you can upload the policy as an update to your Culture Guide or Employee Handbook and send
9:49
notices to your entire workforce requesting their review and e-signature. Make sure everyone knows
9:56
that if they use AI and it makes a mistake or goes poorly, it isn’t the computer program that is on
10:02
the hook—it’s the human. Remember—your role is as strategic as you make it!