(this is an image-heavy edition - make sure that you permit your email client to display them)

Dear Reader,

If you thought to yourself:

...you'd be wrong - I've also been using it to create Studio Ghibli-esque images and all sorts.

​

​

In case you missed it, last week OpenAI unveiled a significant upgrade to ChatGPT's image generation capabilities, integrating the GPT-4o model to create and modify images directly within the chatbot. But is it worth the hype?

Let's find out.

What you need to know about 4o Announced on March 25, 2025, this new feature allows users to generate detailed and accurate visuals through natural conversation, excelling at rendering text, following prompts precisely, and leveraging GPT-4o's extensive knowledge base. ​ Read more

The Evolution You'll Actually Notice

The first thing that struck me while testing this new system was how substantial and immediately apparent the improvements are:

• More photorealistic quality: The photorealism - when it works - is damned impressive compared to the illustration-style results from previous versions
• Accurate text rendering: The system handles text incorporation with impressive accuracy, following instructions with much greater fidelity
• Complex compositions: Multi-element scenes with detailed specifications are handled with better precision
• Improved contextual understanding: The generator better interprets nuanced prompts, reducing the need for multiple attempts

I've noticed it's particularly good at transforming existing images, though it's still catching up to creating new ones from scratch. The Studio Ghibli filter is surprisingly well-executed and has become very popular among users already, even if OpenAI is wading into some murky IP waters. And speaking of which...

Try Not to Say 'Muppet'

Or Simpsons, Marvel, or anything else with a specific style that infringes on intellectual property. You can try, and it might get you so far before it cuts out. Instead, it's better to take your idea and make it more generic:

"Please recreate this photo using felt-style puppet characters that resemble a 1970s variety show vibe – think big eyes, fuzzy textures, and playful personalities." works pretty well instead.

Real-World Time Savers: Practical Applications

Streamlining Product Development

Product teams can now rapidly create concept visualisations for packaging, displays, or the products themselves. This capability allows for quick iteration and feedback collection without waiting for specialised design resources, potentially reducing the concept-to-approval timeline from weeks to days.

The ability to specify exact colours, text placement, and product details means these visualisations can closely match brand guidelines and product specifications from the very first draft.

Creating Educational Materials More Efficiently

The system's improved text handling makes it particularly valuable for creating educational content. Custom diagrams, labeled illustrations, and concept visualisations that would typically require significant time in design software can now be generated through simple text prompts.

For educators and trainers without design backgrounds, this transforms what would be hours of work into a straightforward conversation with ChatGPT, making custom visual content creation more accessible and efficient.

Building Website Assets That Match Brand Guidelines

Marketing teams and small business owners can now generate website banners, social media graphics, and other visual assets that conform to specific brand requirements. By including exact colour codes, font styles, and composition preferences in your prompts, you can create consistent visual materials without specialised design skills. You can even throw a screenshot of your existing colour palette at it and get results:

This capability helps maintain visual consistency across digital platforms while significantly reducing the production time for marketing materials.

Current Limitations Worth Knowing About

Understanding the system's constraints helps set realistic expectations for your projects:

• Portrait specificity: The system still struggles with generating specific faces or exact likenesses
• Generation time: Images take about 30-60 seconds to generate, which is longer than some specialised platforms
• Occasional composition issues: Some elements might appear slightly different than described, sometimes requiring additional prompt refinement
• Complex lighting effects: Specific lighting scenarios like backlighting may not always render as expected

I've also noticed it's still catching up to itself in terms of knowing what's allowed and what's not. Sometimes it refuses to create perfectly acceptable images because it's being overly cautious, while other times it might generate something you didn't expect.

Ethical Considerations That Matter

As these tools become integrated into our workflows, they raise important questions worth considering:

Supporting Creative Industries

While these tools can dramatically speed up certain creative tasks, they work best when viewed as collaborative tools that enhance human creativity rather than replace it. Consider using AI-generated images as concept starters, mood boards, or draft visualisations that can then be refined or reimagined by professional creatives.

This approach leverages the speed and versatility of AI while still valuing the unique perspective and skills that human creators bring to projects. I lump the widespread doom-saying of AI tools in the same camp as those who bemoaned the advent of the Internet and likewise Wikipedia. These are tools meant to be used to help us get better at what we do, not replace what we do.

Transparency in Media

As AI-generated images become increasingly realistic, transparency about their origin becomes more important. When using these visuals in public-facing content, consider adding simple disclosures. This transparency helps maintain audience trust while still benefiting from the technology's capabilities.

Responsible Use Guidelines

Developing personal or organisational guidelines for appropriate use of AI-generated imagery can help navigate potential ethical questions. Consider factors like:

• When to disclose AI assistance in creative work
• Appropriate contexts for using AI-generated imagery
• How to properly attribute the role of AI in your creative process

These considerations help ensure the technology enhances rather than complicates the creative landscape.

🛠️ Get My Agents for Creators, Builders and Doers. My AI Creator’s Toolkit is a growing collection of lightweight, task-focused AI agents designed to help you get small jobs done faster. No fluff, no jargon, and no need to learn prompt engineering (but if you wanted to, there's also a tool for that!). Here are a few of the tools included: Jeannie – helps you come up with ideas for what kind of GPTs you could build Jamie – a YouTube assistant that helps you plan, script, and optimise your videos Webster – turns your meeting notes into clear actions and takeaways Seymour – quickly generates SEO-friendly meta descriptions from any webpage Prompto – takes a messy idea and turns it into a clean, usable GPT prompt The AI Creator’s Toolkit is all about cutting out repetitive tasks and giving you a shortcut to useful results. More tools are being added every month, so sign up today! Check out the AI Toolkit 👉 You can try it for free (includes 10 credits)

​

Getting Started Today

Everyone has access now, including free users. To begin creating images, simply describe what you want in clear, detailed language. The more specific your instructions regarding composition, style, colours, and text elements, the better your results will likely be.

Start with simpler compositions to get a feel for the system's capabilities, then gradually explore more complex requests as you become familiar with how prompting affects the output.

For those of you who, like myself, are looking to build toolkit solutions on top of this, we'll have to wait a little bit longer until API access is rolled out.

Practical Power at Your Fingertips

The new ChatGPT image generator represents a significant step forward in making high-quality visual creation accessible to everyone, regardless of design experience. By reducing the technical barriers to creating professional-looking visuals, it empowers users to bring their ideas to life more quickly and efficiently.

Look, I'm going to keep banging the drum - this is not going to replace the depth of skill and creativity that professional designers bring to complex projects. This tool provides a valuable resource for rapid visualisation, concept development, and routine creative tasks--saving you time and energy while expanding what's possible in your daily work.

Until next time,

Jim

​

​​​​

Made with ❤️ in Valencia by Jim Christian. For feedback, please reach out to hello@jimchristian.net.

​