Skip to the main content.
Get Started
Get Started

3 min read

Reinventing Graphic Design with 4o Image Generation

Reinventing Graphic Design with 4o Image Generation
Reinventing Graphic Design with 4o Image Generation
5:53

What is GPT-4o Image Generation?

GPT-4o’s image generation is a feature in ChatGPT that lets AI create images from text prompts, built right into the platform for a conversational, intuitive experience.

It’s “omnimodal”, meaning it learned from both text and images together, so it understands how words and visuals relate. Beyond just creating photorealistic or stylised images, it can also handle text rendering, follow multi-step instructions, maintain context across multiple prompts, and apply general world knowledge.

The use cases are wide-ranging: from fun, creative projects like turning photos into anime-style characters, to business applications like logo creation, product mockups, or even generating different views of a 3D model for mesh development.

To explore its potential, we used our free eBook, HubSpot Content Hub Overview Guide, and created a simple, brand-aligned mockup, which we then turned into a social post by combining it with other AI tools.

HubSpot Content Hub Overview Guide mockup

 

What Makes 4o Image Generation Special?

It's not just about new art styles. GPT-4o dramatically improves on older AI tools when it comes to embedding clear, accurate text into images, such as signage or logos.

It also handles complex prompts involving multiple objects and maintains consistent details across a series of related images.

Here's an example: 

  1. Take a picture of your clothing
  2. Create an Al model wearing those items
  3. Make the Al separate those clothing items as a mockup
  4. Then have the Al create a sample website design from the set of images that you've just generated

Website design mockup

 

You'll only need a prompt, maybe a reference image, and your imagination. You can ask for an image and then tweak it with follow-up requests in the same chat.

It also appears to take a different approach than older tools like DALL-E 3, potentially generating visuals in a more step-by-step, structured manner.

A whimsical city street scene at night, featuring a red-haired woman in a velvet cloak haggling with a steampunk vendor.

Dall-E 3 version

A whimsical city street scene at night, featuring a red-haired woman in a velvet cloak haggling with a steampunk vendor.Using the same prompt in 4o Image Gen version and adding a touch of realism

 

How Is The Industry Reacting?

People are genuinely amazed by the quality and ease of use. Many are discovering creative and commercial possibilities, especially when these tools are combined with others like Sora, Kling, and more to add animation and motion.

For example, here's what we got on our first attempt using only our HubSpot Content Hub Overview Guide as a starting point (without any editing):

HubSpot Content Hub Overview Guide

 

Next, we created a graphic and added the text "GRAB YOUR FREE COPY NOW". Then, we used an AI tool called Kling to turn it into an Instagram story video.

Prompt for video in Kling: a 3D 360 camera shot of a magazine floating in the air with elements floating around it, dark and fuzzy purple background, in 3D cinematic.

 

Then, using ChatGPT 4o again, we prompted it to have a model holding our product:

A warm, well-lit indoor portrait of a smiling Caucasian man in his 30s to 40s wearing a grey hoodie and holding up the booklet. The room behind him is softly blurred with neutral tones. Prompt: A warm, well-lit indoor portrait of a smiling Caucasian man in his 30s to 40s wearing a grey hoodie and holding up the booklet. The room behind him is softly blurred with neutral tones.

 

We also tried to create a shirt out of it, have the model wear the shirt, and used Kling to animate the image:

Prompt for video in Kling: slowly zooming out shot of the man wearing a t-shirt showcasing his new t-shirt and looking at it in his living room shot with an iPhone camera

 

Lastly, we brought the previous image of the model holding our product to life. Using Kling, we animated him to speak a custom script we prepared, adding a dynamic, user-generated content feel to the visual:

Prompt for video in Kling: the camera zooms out,as the smiling guy holds the magazine. 

 

Reinventing Graphic Design

“Prompt engineering” is quickly becoming a real career path, one that might seem niche now but could grow in demand (like early computer programming or being a YouTuber).

While some designers and artists are concerned about disruption, many see this technology as a creative partner, much like when computers first entered the design world. AI helps eliminate repetitive tasks, freeing up time for more strategic and imaginative work. It can make professionals more productive and potentially increase the demand for design overall.

The same trend is playing out in software development. AI is boosting coder productivity, and while that may affect some IT services, many believe it will lead to the creation of more software, not less. (A real-world example of Jevons Paradox, where increased efficiency leads to higher total output.)

These tools also make content creation more accessible. While some fear that this lowers the bar, it can also democratise creativity, giving more people the tools to express ideas. OpenAI’s CEO Sam Altman believes this is a net positive for society.

For those working in companies that are slow to adopt AI, Sam Altman suggests moving to faster-moving companies to learn and stay relevant, as non-adopting companies may struggle in the long term.

The creative role is evolving: from doing the work manually to directing the AI, managing context, ensuring quality, and applying human taste. While AI might automate some tasks, the need for good design and strategy remains strong.

 

A Positive Outlook

Prompt for video in Kling: slow-motion sequence captures the conversation of an old man on his telephone and a girl wearing red hood talking to each other, people walking the background at evening, as the camera zooms out

 

Sam Altman notes that while AI is a "crazy powerful tool," its societal impact so far has been less disruptive than sci-fi predictions, something he views as a good sign.

At its best, AI is here to help people do less repetitive work and spend more time on meaningful, creative projects.

The question is: Are you already adapting?

 

Read more related blogs

 

GPT-4o: A Leap Forward in AI Accessibility and Multimodal Interaction

GPT-4o: A Leap Forward in AI Accessibility and Multimodal Interaction

OpenAI, unveiled GPT-4o (“o” for “omni”), their latest flagship model designed to bring GPT-4 level intelligence to all users, including those on the...

Read More
Claude 3.5 Sonnet: An Enhanced Chatbot Experience Compared to GPT-4o

Claude 3.5 Sonnet: An Enhanced Chatbot Experience Compared to GPT-4o

Anthropic has unveiled Claude 3.5 Sonnet, announced on June 21, 2024. This release is a major leap in AI, outperforming both competitor models and...

Read More
Testing and Comparing Free Versions of Claude, Gemini, and ChatGPT

Testing and Comparing Free Versions of Claude, Gemini, and ChatGPT

Large Language Models (LLMs) have rapidly advanced in recent years, demonstrating remarkable capabilities in text generation, problem-solving, and...

Read More