You're the boss of AI — act like it.
First published on LinkedIn on February 13, 2023.
There are so many people talking so much about AI, it reminds me a little bit of the last "technology transformation" we had when social media suddenly became a real channel years ago.
At that time, I was thrust into the ethics of brands and social media as a result of suddenly having two-way conversations in front of vast audiences about the ethics of paying influencers well before there really was such a thing.
So today, as AI "Experts" emerge, and as news coverage abounds in terms of the opportunities and potential pitfalls that Artificial Intelligence tools open to humanity, I pose the following:
When it comes to the use of Artificial Intelligence, you're in charge. That means you're responsible.
My friend Toby sent me an incredibly compelling article the other day from The Atlantic, "Talking to AI Might Be the Most Important Skill of This Century," and I think it's true.
But before we train, it's important to realize that we're all just finding our way... These systems themselves are built to learn from the masses, so we're all learning and teaching as we go. And few of us are deeply considering the consequences. Yet.
Guidelines and Guardrails 1.0
Guidelines and guardrails are beginning to emerge. Enter, Department of Commerce's Artificial Intelligence Risk Management Framework. In it, there are plenty of descriptions of risks to people, organizations and our societal ecosystems called out:
"AI systems, for example, may be trained on data that can change over time, sometimes significantly and unexpectedly, affecting system functionality and trustworthiness in ways that are hard to understand. AI systems and the contexts in which they are deployed are frequently complex, making it difficult to detect and respond to failures when they occur. AI systems are inherently socio-technical in nature, meaning they are influenced by societal dynamics and human behavior." - Executive Summary
The Framework (as only a government document can) goes on to further define risks on spectrums of the stage of AI ecosystem (we're in the very early stages of widespread adoption - AKA the Technology Trigger in Gartner's Hype Cycle) and it also defines each and every characteristic of trustworthiness' "to-be" not "as-is" state:
"Characteristics of trustworthy AI systems include: valid and reliable, safe, secure and resilient, accountable and transparent, explainable and interpretable, privacy-enhanced, and fair with harmful bias managed." - p. 12
But getting from the "as-is" to "to-be" desired state requires thoughtful deliberation, a sense of your own implicit bias, and a big sense of accountability as you use these systems.
More on that in this space. For sure.
Practical Advice for Marketers 1.0
Lest you think I'm just here to complain, let me share some advice I've gotten from my afore-mentioned friend, Toby Graham, as I asked her to share her use of AI tools to help her do her job. She and my engineering friend Eric both responded to my requests for some advice. (Eric's advice and experiences with AI are upcoming, in the next blog post.)
JLJ: How are you using AI to help you do your work?
TG: Copywriting has never been my strong suit, which is funny for someone who’s in charge of content strategy. But I make up for it with my editing skills. I can visualize the story I want to tell and identify what needs to be included so that my message resonates. But I come up short crafting the words themselves.
AI helps breathe life into my content by taking my rough outlines and turning them into richly detailed pieces that drive home their points effectively.
Think of laying down stepping stones vs. paving a sidewalk. In this case though, the end result is a polished story rather than simply a path from one point to another. With the help of AI, I can lay out the basic components of my piece, and it will do the hard work for me, providing more descriptive language and relevant facts to enrich my story.
JLJ: What’s the biggest benefit of using AI for what you do?
TG: I hate to say it, but I’m leaning on AI in the place of my copywriting freelancers. Not that I don’t respect their craft.
I'm always looking for ways to streamline my content calendar, and it turns out AI can really help with that. Whether I have a rough draft of an idea or just bullet points on what I'm trying to create, AI does the heavy lifting in turning them into compelling narratives without having to sink time coordinating with freelancers or developing creative briefs.
AI helps me take a messy outline of a piece I want to develop and craft a compelling narrative on my timeline, not someone else’s.
JLJ: Given your experience, what’s the biggest ‘gotcha’ that you should look out for?
TG: I haven’t come across gotchas yet, but I certainly wouldn’t take AI output and plop it straight into a piece without some finessing. This finessing can help you stay away from potential plagiarism issues as well. AI output tends to be wordy right out of the gate and needs some smoothing to be usable.
JLJ: What advice would you give to someone who does what you do in the use of AI?
TG: Build in some futz time.
I’m finding that to use AI successfully, you need to understand how to craft commands that will trigger accurate and helpful output. Think of this like learning American Sign Language - a completely different form of communication that unlocks a whole new world.
JLJ: Did you use AI to answer these questions?
TG: Aaaannnd…I’m not going to tell you if I used AI to answer these questions.
If you find this piece poorly written and disagree with my opinions, it’s 100% robot generated.
If you think it’s an effective piece that gets my point across well…then I take all the credit for it.
Ha! Well done, and thank you Toby! The credit is all yours. Thank you, Eric, in advance. I'm looking forward to more highly interesting conversations and work in this space. And PLEASE feel welcome to connect and comment as your experience with / or questions about... the implications of AI grow. Because they will. ;-)