I just returned from speaking at a conference for professional communicators where, once again, AI and ChatGTP were hot topics. While there were many presenters talking about how to use AI to save time while levelling up their communications, there was little discussion about responsible use of the technology.
As someone who has worked in public relations and corporate communications for over 25 years, I’ve seen and embraced many new technologies over the course of my career. Whether it’s use of video, social media or AI, a common trend I’ve noticed is the lag time between technology adoption and consistent policies and practices in organizations around its usage.
The following are some areas to consider in developing much needed AI policies and/or guidelines.
Wild west
Watching the usage of AI and ChatGTP in organizations, it’s pretty clear this is being done with little direction. With leaders struggling to catch up to adopters, we’re seeing a whack a mole approach to AI.
Some leaders see the benefits of this technology, while others remain unconvinced, trying to hold back the inevitable. The result – unclear and inconsistent usage of AI with a lack of understanding amongst staff.
Regardless of which camp an organization falls under, many are still trying to figure out how to establish clear perimeters of acceptable usage.
Privacy protection
While there are many usages that need to be considered, one that requires immediate attention involves protecting the privacy of staff, customers, and confidential information.
The downside of the human-like interaction of ChatGTP is people tend to forget they’re engaging in a public platform. This often results in oversharing sensitive or personal information, assuming the conversation is confidential.
It puts at risk proprietary or confidential information being shared in a public setting. While the employee may think twice of doing this in a community meeting, they might not be as cautious when typing into ChatGTP on their laptop.
With many people not truly understanding how ChatGTP and other AI features work, it’s important to go beyond a usage policy and take the time to train staff. Just as staff go through cybersecurity training, so they don’t click on unknown links, AI training is also essential to protect privacy and confidential information while leveraging its benefits.
Social media
Another area where I see a lot of reliance on AI is in social media posts. You can quickly spot the organizations who have turned over the reins of their content development to AI. These are the posts with lots of emojis, heavy on copy, and lacking any personality. All the posts look and read the same.
Many people who use AI to create their posts don’t understand the downside.
What they don’t know is that these posts are often flagged by the social media platforms and are being placed at the bottom of the algorithm. The other challenge – the public can spot these a mile away. In turn, the trust and integrity of your brand decreases in the eyes of your community.
So, what was originally a good idea to save staff time is actually hurting your organization in terms of reach and community perception. You’re better off not posting at all vs letting AI generate your social media content. It’s fine to use it as a starting point to get the creative juices flowing, but don’t let AI be your social media specialist.
Another key education opportunity for all staff.
Attribution
While there are definitely many great reasons to use AI, I would encourage your organization to determine when should it be disclosed.
For example – I used ChatGTP to help write a headline for this article and check my grammar. I don’t feel that needs to be disclosed. However, if I entered a bunch of information into ChatGTP and let it write this article for me, it’s no longer original content. That I feel deserves to be disclosed – especially if I’m writing the article on behalf of a paying customer.
What disclosure and attribution looks like depends on the work being done, the organization and the industry. But it’s a key area for consideration.
Crisis communications
This is a usage that is starting to gain momentum, and as a crisis communications expert, scares the crap out of me. I’ve even heard professional communicators admit to using AI to help draft press releases or create their crisis communications plans.
Stop!!! This is a HUGE risk.
Why? Because every crisis or issues management event involves real people, confidential information and internal processes that need to be safeguarded and kept internal.
Using AI to help draft messaging or create strategies is a huge breach of privacy.
My advice – pick up the phone and hire a crisis communications consultant instead of relying on a public platform for guidance.
Duty to protect
This leads me to my final point for consideration in your AI policy/guidelines – not entering names into AI. Again, the information you’ve entered will continue to be connected to the individual.
Many AI tools do not differentiate between validated and unvalidated information.
A best practice is to keep names, or any information that can be directly connect to them, private.
Clear perimeters
If there’s one takeaway that I want you to leave with, it’s this – don’t wait to create policy and/or guidelines for AI usage in your organization. And make sure you educate your staff along the way.
Like any technology, the sophistication and usage of AI is always changing. What’s important is your organization is being proactive in changing with it.
While it would be great to have a standardized AI policy and/or guidelines that I could share with you, this just isn’t possible. Each organization, and even departments within an organization, will have unique needs, limitations and opportunities.
I encourage you to start putting some guidance in place for your staff before you’re confronted with a negative situation that could’ve been avoided.