...
test
AI-Risks-Blog-Header-Image
randi square headshot Randi Anderson

Because Brand Strategy Isn’t a Prompt

AI is such a powerful tool, it seems able to do anything—like it’s our technological fairy godmother. Honestly, we know from experience that it can be a godsend for making daily tasks a little faster and easier! But like any technology, it definitely has its limits. There are still things you should trust to the human experts, like financial planning and medical care (we’d like to see ChatGPT’s clinical experience!). And, yes, that list includes marketing and branding advice.

You can probably guess what’s wrong with trusting Dr. Gemini. But when you’re working on your company’s new brand, what’s the harm? To explain, we’ll need to start by busting a myth.

The Illusion of Objectivity

Seeking advice from AI chatbots can seem like a sound choice precisely because they’re machines, and we imagine they will be more objective and fact-based than a human. Unfortunately, recent research has shown that AI is not nearly so objective as we think!

For one thing, as Stanford found, these chatbots have a strong bias toward likability. The very process of developing and training AI encourages it to be your personal yes-man, giving you answers it anticipates you will like. The goal is your immediate satisfaction rather than giving the whole truth or meeting your goals long-term. But why does this happen?

When large-language models (LLMs) like ChatGPT or Gemini are trained, there are three basic phases:

  1. Studying, when they’re fed a ton of information that they analyze, like a student cramming for a test.
  2. Learning to answer questions, when humans train the AI model to give appropriate responses.
  3. Reinforcement learning from human feedback (RLHF), when humans train the AI model to give “preferred responses” based on scores and rewards.

Researchers have noticed the third stage, RLHF, has been causing problems. In that stage, LLMs are trained to maximize human satisfaction, which sometimes conflicts with truthfulness. Princeton researchers developed a “bullshit index” (their words!) to measure the confidence of AI’s answers against their accuracy. They found that after RLHF training, unverified claims got worse, and instances of “paltering” nearly doubled: that’s when AI tells you something that is technically true, but only because it has left out some important bits!

In other words, AI is trained to answer like a nervous employee to its boss: if it doesn’t know or doesn’t think you’ll like the answer, it might fluff it.

So if you plug your human-made logo into AI and ask it to judge how effective or “good” it is, it’s going to give you an answer whether that answer has any sound basis or not. In fact, AI will always suggest changes, even if you don’t need them! It will never tell you something is perfect (you can test this!) because it’s programmed to refine, even when refinement isn’t necessary or in your best interest. It will even endlessly critique its own suggestions.

AI Is Not the Branding Expert

When you want to know how effective a logo or tagline would be for your company, asking AI feels like tapping into a vast store of data. We’ve already seen that the bot’s answers can easily be fluffed. But we also need to look at the data itself: AI is pulling together ideas from countless sources, and not all of them agree. They’re not all true thought leaders, either.

All these voices can cloud the message with their conflicting approaches. Keep in mind why you hired your brand designer in the first place. We hope it’s because their brand philosophy—like ours at KickCharge—has a proven track record. Why throw in the opinions of those who don’t share that philosophy?

Consider this, too: AI doesn’t know your ideal customer, your target market or your specific competitors. It’s powerful, but not powerful enough yet to zero in on what will work in your unique situation.

Your logo designers and copywriters, on the other hand, have dived deep into your client brief and competition and can craft a brand tailored to disrupt your market and make you stand out. They have the real-life, in-depth experience to evaluate whether a brand asset (a logo, color palette, or name, for example) will remain distinctive and build equity for you in the long run.

Simply put—anticipating what will and won’t work is still an area where human expertise and intuition are your best bet!

AI Logos: The Three-Eared Elephant in the Room

Logo creation is another arena in which AI’s abilities fall short. We all know about AI’s infamous hallucinations—statements that are completely untrue, images that don’t make sense. The technology is evolving rapidly, so we’re less likely to see an elephant with three ears, and sometimes hands even have the right number of fingers. On the other hand, the weird stuff still happens far too often.

In logo design, you have not only the hallucinations to contend with, but also problems specific to good branding and practicalities.

Let’s start with branding. At KickCharge, our philosophy is disruption—looking different and standing out. As we always say, you want to be branded, not blanded! By its nature, AI generates images based on what is already out there. You’re getting something derivative rather than unique.

It’s true that a highly trained illustrator has the expertise and trade lingo to craft a detailed prompt and get something closer to a unique vision… but let’s face it, most of us don’t have that skill. Think of it like this: a seasoned doctor is more likely to get a good answer on a medical question because he or she can write a prompt informed by at least a decade of training and experience. Like any trade, brand design has its own language and technical ins and outs!

For Consistent & Effective Branding, You Need a Human

If you do manage to get a unique logo, you have one more hurdle: making it scalable. Your branding needs to be consistent, and in order to keep it looking great across all platforms—from your website to your print collateral to your vehicle wraps—you need the ability to make your logo bigger or smaller without losing image quality. A blurry, pixelated image just looks unprofessional.

To make your logo infinitely scalable, our brand designers do something called vectoring. It’s a time-consuming process that AI hasn’t mastered, and while theoretically you can vectorize an AI-generated image, it is far more difficult than one specially designed by an experienced illustrator. The reason? AI visuals tend to have complex lighting effects and contours that are tough (sometimes impossible) to render in vector.

Ultimately, this means AI might be a good place to get inspiration, but using it in the actual logo development can significantly complicate and drag out the process.

The Right Way to Use AI in Brand Decision-Making

Does this mean you should never use AI while working with us on your brand? Of course not! It is still a highly useful tool when used effectively. Here are some things to keep in mind:

What to Avoid

  • Using “leading language” in your prompts. The way you word your AI prompts makes a big difference in the answer you get. For example, if you ask, “Show me what’s wrong with this logo,” you’re giving it a huge bias right out of the gate!
  • Asking AI to judge the marketing effectiveness of different options. As mentioned above, the chatbot doesn’t know your specific market or your competitors. On top of that, it is likely to fluff an answer just to satisfy your expectations.
  • Giving your designer an AI-generated image to render as-is. Even if it’s usable for your brand (a big “if”), the image could add a lot of billable hours through the reworking and vectoring process.

What Could Help

  • Providing plenty of context and carefully wording your prompts. The more specific details and context you can offer when asking for suggestions, the more targeted the chatbot’s response will be. At the same time, be careful that your language is as neutral as possible and doesn’t steer the bot toward a particular conclusion. This can take a lot of trial and error.
  • Using AI to express something you had in mind visually. Sometimes you want a certain color or a particular tweak made to a logo, but you don’t know how to express it clearly. In that case, a carefully crafted AI prompt might visualize what you’re trying to communicate and help you and the designer skip a little back-and-forth.

Finally, keep in mind that if you’re feeling unsure—if there’s something bothering you, and you don’t know what—it might be because the whole thing is new. It’s different, and it’s your company. That can be unnerving! Still, unless you’re using AI to communicate something highly specific, reaching for it in those moments can blur the strategy and create setbacks instead of forward momentum.

The best thing to do? Express any concerns you might have to the team working on your brand, and trust the process. Your human designers know what makes a good, clear visual as well as an appropriately disruptive brand. After all, this is what we’ve been doing for home service companies for over 30 years. It’s what we’re all about!

eco vehicle wrap view of the back paneling and passenger side