Insights

Curated Creativity: How Chatbots Will Change Web Design

December 22, 2022
Blog feature photos gpt

As chatbots get smarter and and can tackle more tasks, they'll play a bigger role in our daily lives, managing many of the routine tasks and decisions that we handle ourselves now. The creative process will be drastically impacted, too, not just because it's tough for creatives (which it is), but because it's practically impossible for non-creatives.

Instead of starting with a blank canvas, maybe the app will ask for your current website URL, any content you've already written, or general idea of what you're looking for. From there, you'll get design suggestions in the form of a layout to start filling in.

With the click of a button, a first draft of copy will fill a given block on the page. As you adjust the tone of one section, the rest will adjust accordingly. Photorealistic imagery with lifelike human models and representations of your own products in action will be generated on the fly based on a training set of source images you've provided. Your visual brand, fonts, and colors will be applied to text styles, buttons, and links.

A web design GPT will apply convention, consistency, and randomness dramatically faster than it's human counterparts.

The point is, unlike the custom templates of old, GPT will enable a fluid creative process bolstered by machine assistance. Just as ChatGPT can write the next sentence of a story, and Github Copilot can finish out a function, a web design GPT will apply convention, consistency, and randomness dramatically faster than its human counterparts.

Your Eye For Detail, Encoded

Image generators like DALL-E, Stable Diffusion, and Midjourney are showing they can do some really impressive work with the right prompts. If you add custom training to these models, they can be fine-tuned to do even more specialized tasks. Prompt engineering and custom model training could become a new cornerstone of the creative professional's career.

Think about it: During the design research process, we gather screenshots and snippets of things we find particularly interesting, effective, and delightful. In many instances, we leave notes along with our findings:

"This headline animation is great. Such a smooth transition."

"Love the clean, straight edges in this layout."

"The letter-spaced all-caps label here is perfect for this upscale business."

GPT will be able to process your notes in combination with your screenshots – not only what you like, but why you like it. On top of the library of textures, design components, stock photos, and fonts that designers collect today, a continually refined model – akin to your personal tastes, tone, and preferences – will form a training data set. Download the latest version of GPT, fine-tune it with these personal preferences, and you'll have a creative sidekick that is uniquely you.

The library of textures, design components, stock photos, and fonts that designers collect today will become a continually refined A.I. model – akin to your personal tastes, tone, and preferences.

Intellectual Property and Privacy

Of course the same technology will apply to artists, writers, and architects. It will also be used in countless personal ways, like generating an auto-draft of an email in your unique voice or creating video blogs using a deep fake video avatar indistinguishable from your living self. While that avatar will nail every cut on the first take, privacy concerns will abound. If your creative work (or the personalized engine behind it) isn't encrypted or tagged with a verifiable personal signature, it could create all sorts of unforeseen problems.

The same will go for businesses that want to leverage A.I. across the whole of their intellectual property and private internal knowledge base. Concurrent advancements in identity management, privacy, licensing, and security will be crucial when it comes to fully capitalizing on the technology.

Food for Thought

  • In the same way a result on the first page of Google is valuable, figuring out how to own the position of the default suggestion could be financially lucrative for early adopters. It's worth considering how early winners of organic search won and why. The first wave of content strategy came from small niche bloggers pursuing their passions, not corporate strategists.

  • Considering your personal or business chatbot will only be as personalized as you train it to be, consider exploring the field of Personal Knowledge Management (PKM) to give yourself a head start in making your chatbot more like you from the start. Well-structured content management systems, apps like Obsidian, and concepts like Building a Second Brain are likely to be worthy time investments.

  • As GPT technology advances, it will become increasingly important for businesses to have a strong online presence on which to optimize their customer-facing operations with machine learning. In a previous article, Business Chatbots and The Race to Better Web Content, we talked about the same. Those who are best positioned to explore and adapt will have a strong competitive advantage.

A Future of Unparalleled Human Creativity

If GPT is trained on centuries of original human work, it's easy to feel that removing humans from these creative processes could result in a recursive loop of mediocre machine output. But it's far more likely that individual perspective, personal experience, and taste will rise to the top and remain the undeniable essence of creativity. Just as Picasso broke from his classical training to give birth to cubism, the most innovative creatives will push the boundaries of what's possible, leading to entirely new forms of personal expression.

ChatGPT was used throughout this article. It was especially helpful for forming more cohesive paragraphs from first drafts, and fixing up clunky sentence structure and punctuation. The feature image was created by Midjourney using the image from our previous chatbot article as a starting point. Is it perfect? Not yet. But fun to play with? Oh, yes.