However, not all AI tools and results are ideal. While using AI can be convenient and quick, that doesn’t necessarily equate to quality or trustworthiness. AI may not be the best approach for your particular website redesign. But: if you do choose to use AI, take care to avoid these pitfalls as you redesign your website.
Generative Imagery Is Obvious
We’ve all seen the images generated by AI. They’re smooth and seemingly perfect—until you notice seven fingers on one person and strangely misaligned teeth on another. In general, using stock photography and real images on your website will always provide a better human touch than current AI-generated images that can feel sterile and eerily familiar.
Use high-quality images on your website when possible—or your brand perception may suffer.
But:
If you’re hoping to save time with image generation, use AI instead for more abstract ideas that are less likely to have noticeable mistakes. If, for example, your industry is software-based, it may be worth prompting AI to generate general ideas instead of particular product uses.
Additionally, AI can be well-used to enhance existing imagery—to extend a photo, for example, or remove blemishes or replace a background. Instead of starting from scratch with AI-generated images, use it to boost or tailor what you already have.
AI-Generated Designs & UX Don’t Differ
It can feel overwhelming to come up with a new interface. And AI can help—but should it? Vet everything carefully. If you’re claiming something as original, it probably should be. And with AI being fed data from a multitude of sources, it can end up being wildly unoriginal, or possibly even viewed as a “rip-off”.
Instead:
Consider using AI for a particular section to build a design pattern. Then: customize it as much as possible. The end product shouldn’t look like everyone else’s. It should signify what sets your brand, company, and offerings apart. AI results are not final products and should not be treated as such.
If you’re simply doing your research, check out Mobbin. Their library of different UX interfaces can help you determine how best to solve UX issues and keep an eye on trends and solutions for your website.
Machines Learn Based On Data—Proprietary or Not
What data are you handing over to your AI tools? First, is it yours? Second, is it proprietary? Consider that what you’re designing and creating can be used for machine learning. Effectively everything you plug into your AI tools could be drawn out elsewhere by another user.
If you’ve given something over (data-wise) that isn’t yours to share, that’s more than a gray area—it’s a risk! Ensure that everything you’re feeding into the machine is information you have permission to share.
Generic Asks, Generic Results
When you’re using AI, being more specific can get you more specific (and more useful) results. If you’re generating images (abstract ones, of course) and you ask for a “ray of light” without specifying size, color, or shape, it may take some time to land on an image that suits what you need.
Or, if you’re trying to customize UX, setting specific parameters can help you eliminate any extra complications in the backend.
Just remember: Generic prompts get generic results. And that doesn’t equate to anything remotely original or interesting. Your website design will start to blend in with the rest, which is the wrong way to go when you need to stand out.
Using Only AI for Coding
Exclusively using AI for anything is, in general, not a good practice. It can come in handy for bits of code (think snippets) but relying fully on AI to create full code for websites could land you in a spot where you can’t decipher which pieces of the code are doing what.
That said, AI can be leveraged for coding for CSS animation or doing direct in-browser designs to determine what’s wrong. It does require some familiarity with code if you plan to use AI this way. Rely on your developers to guide you instead of creating code without background knowledge.
More Tips for Using AI in Website Design
Use Your Wits and Your Team – A good rule of thumb when it comes to using AI in website redesigns: Don’t replace everything you do with AI. It’s an assistive tool and should be tailored and customized by actual humans to ensure it’s effective (and not bad reproductions of what’s already been done). Use your team and your own talents and experience to start—and to finish.
Continue to Refine – Find better ways of asking for things—try new prompts and see what kind of results you can get.
Limit AI to the Tedious Tasks – Use human creativity to set the pace and put the robots in charge of the drudgery…as it should be. Think: Removing backgrounds in photos rather than creating imagery from scratch.
Check Your Work – Don’t take it all at face value. Test the code. Look carefully at the imagery. Determine if that Experimental AI section layout really is as slick as it looks. A good tool here: Stack Overflow—coders help other coders solve coding problems.
And finally:
Treat Each Project Uniquely – There is no one-size-fits-all for website redesign. AI builds things to appeal to and solve the problems of the masses. But your website isn’t for the masses; it’s for your ideal customer. Design it specifically for them.
