Are we finally at the point where we can use AI to generate images for marketing and comms? Not just for fun, not just for messing about, but for actual professional use?
That’s the question I want to answer today, because OpenAI has gifted us a significant update to ChatGPT’s image capabilities (it dropped on December 16th). It’s called GPT Image 1.5, and it’s worth paying attention to.
I’ll give you a prompt to test yourself, and then we’ll tackle the big question everyone asks me in training: who actually owns these images?
What’s Changed?
Three things stand out with this update.
First, speed. OpenAI says it generates images up to four times faster than before. And in my testing, that holds up. You’re not sitting there waiting anymore, it’s near-instant for most prompts.
Second, precision. This is the big one for marketing and comms people. Previous AI image tools had a frustrating habit of ignoring what you asked for. You’d say “change the background” and it would change the person’s face. You’d ask for a small tweak and get a completely different image back.
GPT Image 1.5 is supposed to fix that. It keeps the elements you want, lighting, composition, facial likeness, and only changes what you ask it to change. OpenAI is positioning this as a “creative studio in your pocket” that can handle precise edits, not just generate-and-hope.
Third, text rendering. This has always been a disaster with AI image generators, words coming out garbled, letters backwards, nonsense text. The new model handles text much better. That matters if you’re creating anything with headlines, signage, or calls to action.
There’s also a new dedicated Images tab in ChatGPT with preset styles and filters, which makes it feel less like you’re wrestling with prompts and more like you’re using an actual design tool.
You should test it!
Right, I want you to try this yourself. Here’s the prompt I used:
“Create a professional social media graphic for a government communications team announcing a new public consultation. Include the headline ‘Have Your Say’ in bold text. Use a clean, modern style with a blue and white colour scheme. The image should feel trustworthy and accessible, not corporate or cold.”
Go put that into ChatGPT and see what you get, here’s my result!
So is this model ready for professional use?
Here’s my honest take.
For social media graphics, blog images, presentation visuals, internal comms materials, yes, this is now genuinely useful. The speed means you can iterate quickly. The precision means you’re not starting from scratch every time. And the text rendering means you can actually include headlines that are legible.
For brand-critical work like logos, for anything that needs to be legally protected, for high-stakes external campaigns, I’d still be cautious. And that brings us to the copyright question.
Who owns the graphic?
This is where it gets interesting, and I get asked this constantly in training.
Let me break this into two parts: what OpenAI says, and what the law says. Because they’re not the same thing.
What OpenAI Says
OpenAI’s terms of service are clear: they assign you ownership of what you create. Their exact words are “you own the output” and they transfer any rights they might have to you. So from OpenAI’s perspective, the image is yours. You can use it commercially, you can sell it, you can put it in your marketing materials.
What the Law Says
But, and this is a significant but, OpenAI giving you ownership is not the same as copyright protection.
In the US: The Copyright Office has been consistent. Copyright requires human authorship. If an image is generated entirely by AI, with no substantial human creative input, it can’t be copyrighted. It essentially sits in a kind of public domain.
In the EU and Ireland: The position is similar. EU copyright law requires that a protected work must originate from the author’s own intellectual creation. The Court of Justice has interpreted this to mean a natural person must be responsible for the creative choices. Works generated solely by machines, without significant human intervention, don’t qualify for copyright protection.
The EU AI Act, which came into force in 2024, doesn’t actually change this. It focuses on transparency about training data and labelling AI-generated content, but it doesn’t grant copyright to AI outputs.
What This Means Practically
You can USE ChatGPT images for your work. OpenAI has given you permission.
You can’t PROTECT them. Someone else could take that same image and use it. You couldn’t send them a cease and desist. You couldn’t claim infringement.
For most marketing and comms purposes such as social posts, blog graphics, presentation slides, internal documents, and that’s probably fine. You’re not expecting those to be legally protected assets.
For logos, brand identity, anything you’d want exclusive rights to AI-generated images aren’t the right choice. Not because OpenAI won’t let you, but because the law won’t protect you.
There’s also a secondary risk: the AI might generate something that resembles existing copyrighted work. OpenAI disclaims any warranty about that. If you accidentally create something that looks like someone else’s protected design, that’s on you.
Ultimately…
GPT Image 1.5 is a genuine step forward. It’s faster, more precise, and more useful for everyday marketing and comms work than anything we’ve had before.
But it’s a tool, not a replacement for a designer. And the governance piece, understanding what you can and can’t protect, matters just as much as the capability.
Know it. Use it. Govern it. As always.
About the Podcast Host
Joanne Sweeney is a specialist in public-sector marketing and digital communications. She’s the author of Public Sector Marketing Pro and the creator of the Professional Diploma in AI for Public Sector Marketing. With years of experience in training, consulting, and advocating for digital transformation, Joanne provides insight and guidance for government agencies navigating the digital landscape. She is also the host of the award-winning Public Sector Marketing Show podcast.
