AI Ethics & How Marketers Should Think About AI

Potential AI dystopia? Or just another tool? Art inspired by Gee Vaucher (and likely through copyright infringement)

Alright, folks, let’s dive in. AI is here, it’s evolving, and it’s making a huge impact on how we market to consumers. But before we start patting ourselves on the back for being the smartest marketers in the room, we need to have a serious conversation about the ethical side of using AI in marketing. There are some big risks at play, and ignoring them could cost us more than we’re willing to pay.

The Problem of Artificial Diversity

Let’s start with something that’s been bugging me lately: the idea that AI can “solve” diversity issues in marketing. It’s not that simple. I get it, businesses want to appear more inclusive, more representative of the world we live in. So, the quick fix is to start using AI to generate diverse faces and voices in your marketing materials. But here’s the thing: it’s not that easy. And it’s definitely not that ethical if you’re not walking the walk behind the scenes.

If your workforce is 90% white and you’re using AI to throw a few “diverse” faces into your ads, that’s a problem. We’re talking about digital blackface, yellowface, redface. I’m not exaggerating here, it’s just straight-up inauthentic. People see through that kind of thing. If you want to show diversity, you need to first make real, tangible changes within your company. Show it in your hiring practices, show it in your leadership. Only then can your marketing reflect the diversity you claim to value.

AI can’t fix your lack of diversity. Authenticity does.

AI Bias in the Data

Now, let’s talk about something a bit more dangerous: AI bias. AI learns from data, and guess where most of that data comes from? Us. Humans. That means if the data is skewed, or if it reflects societal biases, AI can pick up on those and, in some cases, amplify them. This is particularly concerning when it comes to things like facial recognition technology. Amazon’s system, for example, which they sold to police departments, is notorious for having a higher error rate when identifying people with darker skin tones.

We need to be really careful about the data we use to train our systems. If it’s flawed, if it’s biased, AI isn’t just going to reflect reality; it’s going to distort it. It’s not just an inconvenience, either. It’s harmful. Think about how much damage can be done if AI leads to misrepresentation, or worse, discrimination. We have to address these biases head-on if we’re going to use AI ethically at all, let alone in marketing.

The Sustainability Problem

AI’s environmental impact is real, and it’s huge; especially when it comes to energy and water. We talk about AI as if it’s this shiny new thing, but we don’t often think about what it’s costing the planet. The carbon footprint is massive, and frankly, it raises the question: are we really willing to destroy the planet just to replace menial tasks?

According to the Washington Post, “writing a single 100-word email in Open AI's ChatGPT is the equivalent of consuming just over one bottle of water.” You can read a synopsis without support WaPo at this link on Green Matters.

Danny Gold, COO of Happy Cabbage, hit me with a sobering reality when I was ranting about AI’s water use. He reminded me that every time I use GPS, it’s like dumping five bottles of water. It’s wild how much we overlook the environmental toll of these technologies. And yes, the solutions exist. We need to start thinking about sustainable water use, recycling water in data centers, tapping into regenerative sources, and cooling systems naturally without relying on carbon-heavy methods. Simple changes, but critical ones.

But here’s the kicker: innovation doesn’t always mean “new.” Take DeepSeek’s approach. They used old chips because, in the race to stay ahead of China, the U.S. only provided China with outdated chips. So, DeepSeek had to get creative and figure out how to make those old chips work more efficiently. And guess what? They’re outpacing the latest tech from OpenAI by a mile, all while spending a fraction of the cost and producing less environmental damage.

So, why is new tech always seen as the answer? Sometimes it’s about working with what we have—getting smarter, not just newer. You don’t always need new, flashier solutions. Instead, we need to think “how can we innovate with what already exists to make AI more sustainable?” That’s a question worth asking.

AI Is Taking, But Not Giving

Is AI really AI after all? Not yet. What we have now that we call AI tools are really just Large Language Models. LLMs are not AI in that they do not have any intelligence to create ideas from. What they do is recognize patterns in language across a shit ton of online content and then repackage it like a high schooler writing an essay trying to avoid plagiarism by rewriting someone else’s content.

In other words, what we call AI is really a massive theft of intellectual property. If you’re using AI to generate content for your marketing, from blog posts to graphics, you’re probably tapping into vast databases of content. But here’s the issue: most of that content is made by someone else, almost always being used without their consent.

Now, I get it. AI is a tool. It helps us scale. But we need to be mindful about where that content is coming from. AI doesn’t respect copyright law unless we, as marketers, make sure we do. I’m not saying AI is inherently bad (far from it). But we can’t just turn a blind eye to the fact that artists, writers, and creators are out there making a living off their intellectual property, and AI could be stealing that from them without paying a dime.

If you're using AI to create content, be careful. Don’t just assume it’s fair game. Check the sources, give credit where it’s due, and make sure you’re not infringing on anyone’s rights. The last thing we need is for marketers to be the ones to start devaluing creativity.

Stop Using Stereotypes

Another thing I see a lot of marketers doing that just drives me crazy is using oversimplified customer personas. I’m talking about lazy things like “soccer moms” and “dad bods.” We’ve all seen them. These are broad stereotypes that have been around forever, and they are exactly that – stereotypes. Sure, they might help you quickly categorize your audience, but they miss the mark on who your actual customers are.

While customer personas aren’t exactly related to AI, how you use AI to write content for you or to create ideas to hit your target customers is very much related to whether you actually did the real work on personas.

The truth is, people are complicated. They’re multifaceted. And when you’re building your customer personas, you need to think beyond simple archetypes. You need to understand their motivations, their struggles, their emotions. If you don’t, you’re just throwing a bunch of data points together and calling it a persona, but you’re not actually getting to the heart of who they are.

To really connect with your audience, you have to empathize with them. Think about acting (I was trained in theatre after all). An actor has to understand their character deeply to portray them authentically. They need to understand their internal motivations psychologically and emotionally. We need to do the same in marketing. Your customer personas should reflect the full spectrum of human behavior, not just a few shallow traits.

Then, when prompting AI, you can be explicit about who you’re talking to and not just add more lazy onto lazy by telling your AI tool to write a blog that will relate to “soccer moms.” Perhaps rethink that moniker entirely, it’s sexist and whitewashing.

AI is Helpful Internally

I won’t lie, AI has huge potential to improve internal communication and efficiency. If you’re using AI-enabled tools like Notion or even simple task management apps, they can really help streamline your workflow and reduce the need for endless meetings and status updates. But, here’s the catch: AI doesn’t replace humans.

At the end of the day, your marketing team still needs to collaborate, brainstorm, and get creative. AI can handle the repetitive stuff, like organizing information or automating certain processes, but when it comes to the human side of marketing (the storytelling, the relationship-building), that’s something AI can’t replicate. Don’t fall into the trap of thinking AI can do everything. Use it where it’s effective, but remember, your team is still your most valuable asset.

Ethical AI Isn’t Optional

Here’s the bottom line: AI is a powerful tool, but if we’re not careful, it can easily backfire. From perpetuating biases to creating fake diversity, AI in marketing comes with its fair share of ethical dilemmas. We need to make sure that as we leverage these technologies, we’re not cutting corners or compromising our integrity.

If we want to use AI to its full potential, we need to approach it with a level of responsibility. This isn’t just about efficiency and automation – it’s about ensuring that the human touch, the empathy, and the authenticity are still there. Because at the end of the day, that’s what truly connects us to our audience.