Did you know AI can now create your marketing images for you? The most exciting benefits of AI creating images for you, rather than you choosing stock photos, is that they can be unique, always use the same model, and be consistent with your brand.
The ideas from this article come from an interview with Alex Asher (CEO of LearnCube) and Rob O’Shaughnessy (expert in AI, AR & VR experiences). In particular, we delve into revolutionary possibilities, practical strategies and easy-to-apply techniques using the most popular AI image generation tool on the market, Midjourney.
Before going forward, why should you care if you manage or market a language or tutoring business? You probably have to pay or find images already for your language teaching materials or for marketing.
There are two amazing promises of AI image generation:
This is a fairly new idea from a practical perspective. As Rob suggests, it feels like AI image generation "has really just come up out of nowhere in the last six months. At this stage, we're at version 5.1 of Midjourney. Many people, even the experts can't really differentiate between, uh, what's a real photograph and what's not."
As a quick anecdote, I was a keynote speaker at the inlingua conference in St Gallen, Switzerland to discuss my book LangTech and also the potential of AI. I only used AI-generated images, with credit to the creator in the bottom left, but didn't tell my audience until the very end... You could see a lot of chin-scratching then when they couldn't believe their eyes.
What's even more remarkable is how accessible this technology is. Many of the services are requiring subscriptions but they are quite affordable. I only pay $12 per month for Midjourney and I've never maxed out my usage.
Before meeting with Rob, I just thought about AI-image generation as being a novel, interesting use of technology but it's only when I thought about what it looks to displace, I got excited.
And what does Midjourney compete with?
Stock photos. Most marketers rely on stock photos because they are relatively affordable, professional and abundantly available. The problems that we know, but tolerate, are that the images are very generic, don't look "real", many companies use the same images and the styles are wildly inconsistent. The marketplace of images is made of the work of tens of thousands of different photographers and artists so there is rarely a consistent style to match your brand story.
As Rob says "now, instead of having stock photos from three different photographers in different scenarios that don't create a good story together" you can create images that have a consistent actor or actress, colour pallet, and style, but are completely unique.
When you think about big brands you might admire like Nike, you can "look at some of their images. Without seeing the logo and just know where that came from... Once you understand color theory and how it's applied, you're never gonna look at another advertisement the same."
The end result of good image prompt engineering is that "it makes it look like you had the budget to go out and do an entire shoot across multiple countries in multiple scenarios with the same actor...now a small brand could realistically have a much more consistent brand profile."
Here are some useful tips for making the most of Midjourney:
Rob explains that these "tools and techniques are just the tip of the iceberg. The fusion of AI and creativity is unlocking potential that many couldn't have imagined just a few years ago. With every new iteration, the quality and realism are getting better. But alongside that, the capabilities of these tools to understand context, emotion, and even nuances are improving. This means richer storytelling, deeper customization, and a more immersive experience for the end-user...However, there's a flip side. As we mentioned, the darker aspects, the ethical concerns, are equally growing. The line between reality and AI-generated content is blurring, and that comes with its own set of challenges... we're still at the early stages of what this technology can achieve. The next year will be about refinement, expansion, and hopefully, a more profound understanding of the responsible usage of these tools."
Before finishing up, it's worth bringing up the ethical (concern about the use of particular artists' work in training) and legal questions surrounding AI-image generation from an ethical perspective. Ethically, there are undeniable concerns regarding the use or imitation of particular artists' works. Legally, the courts will eventually determine if LLMs have used copyrighted images in their training. It will be up to each of us to decide how we use these tools but as of writing, I don't think the risk (tiny in my non-expert opinion) is a significant concern for individuals and SME companies that use AI-generated images. I personally don't think users will be (or should be) on the hook legally for using these publicly available tools. A practical compromise for now might be to exercise caution when prompting the AI, especially when referencing a specific artist's style. That I do think seems particularly questionable.
So hold onto your hats, things are going fast.
If you'd like to learn more about how AI is impacting the language education sector, check out this article on: Why Should Language Businesses Care About ChatGPT? And if you'd like to learn more from Rob, follow his work at Rob Shocks on YouTube. His videos are mind-blowing.