Just as generative artificial intelligence (AI) and large language models have transformed writing, they have also transformed the production of visual materials.

It is now possible to create, edit and enhance videos, images, presentations and other visual materials without needing technical design skills. You can, in effect, create professional-looking visual materials without a designer. However, you do have to know a bit about what you’re doing.

As should be clear from our other pages on generative AI, including Understanding Large Language Models and Using Large Language Models, AI is not a substitute for thought by the creator. In the context of visual materials, it is best considered as a tool to speed up the creation process, and one that needs care in its use. This page explains more.

Ways to Use AI for Visual Materials

First, it is worth considering what we mean by visual materials.

These materials are therefore anything that contains any kind of picture or image, including a moving picture or image. This may be alongside or instead of text.

There are several different ways that generative AI can be used to create, edit and enhance visual materials. These include:

• Text-to-image generation

  There are tools that can convert text descriptions into various different types of images.
  These may include cartoons, infographics, ‘paintings’ or photorealistic pictures. The output very much depends on the input. However, with some practice and experimenting, it is possible to create very good graphics to illustrate articles or presentations from text-based prompts.

  One benefit of using AI-based image generators is that they speed up the process of generating and sifting ideas. You can see a lot of images in a very short space of time, and can also amend the prompt to add or subtract certain aspects. An image generator may also throw up some unexpected ideas, which in turn can suggest new avenues for development.

  There is more about this in our guest post on The Role of AI Image Generators in Creative Development.

• Image editing and manipulation

  Image editing and manipulation have been around for a long time—a lot longer than many of the AI tools. ‘Photoshopping’ entered the language a long time ago as a verb for changing a picture, often in a misleading way (for example, changing the face of the person within it).

  However, AI has made image and video editing a lot more seamless.

  Image editing now includes extending or imputing image backgrounds to cover more space (for example, to change a picture from portrait to landscape or vice versa), removing or adding objects, and recolouring graphics. It is also a lot harder to see when an image has been manipulated, meaning that images and videos can rapidly be used to mislead.

• Video generation and editing

  There are various platforms that allow users to create and edit videos.
  For example, HeyGen allows you to create a short informational video that is a bit like an animated presentation using an avatar. It supplies backgrounds and avatars, and you can choose the avatar and the voice from a huge range. A basic account is free, although obviously what you can do is fairly limited. You have to supply the content and whittle it down to the length permitted. As usual, the output depends on the input, and these tools are very definitely not a substitute for your own thinking. However, they do mean that you don’t have to appear in a video yourself, or spend hours stitching together different ‘takes’ of the same scene.

  You can see some early videos generated using HeyGen on our YouTube channel.

• Creating presentations

  There are various tools available to help you to create complete presentations.
  They include the provision of slide decks and design layouts. These include Canva (Magic Studio). You can create or supply an outline, and the tool will do the rest. The advantage here is being able to avoid ‘death by PowerPoint’ because the tool will help you to design a more visually-appealing presentation.

  You can also use AI-based tools to help you develop images and visuals to include in a presentation. For example, they can analyse data and create graphs and other graphics to highlight your key points.

  Finally, you can even integrate AI-based voice-activated tools to improve the transition between slides.

  There is more about using AI-based tools in presentations in our guest posts, Presentation Skills: Using AI to Help You Thrive in Your Presentations and How Can AI Help in Creating Winning Presentations?

• Vector and graphic design

  Finally, tools like Adobe Illustrator will create vector graphics (editable pictures) from text-based prompts. You can therefore create and edit diagrams, which is useful if you want to illustrate a paper with an infographic or similar.

  There is more about some of these uses in our page on Identifying and Using AI Images.

---

---

General Benefits of Using AI for Visual Materials

Perhaps the biggest benefit—and also possibly the biggest downside—of using AI to create and edit visual materials is that it has ‘democratised’ image creation.

AI-based tools have not so much changed what can actually be done as the ability to make it happen without technical design skills. Instead of needing a designer to create or edit an image, or having to spend time learning to use a very technical program, you can now use a simple AI interface to access some very sophisticated tools.

This is great for ‘ordinary’ users, who can now generate images themselves, but is it good for designers too?

Yes, in many ways it is. They no longer have to do the ‘grunt work’ associated with basic design in the past, such as changing backgrounds and tweaking layouts endlessly in response to client demands. Instead, they can provide more expert input on aspects such as whether the proposed solution actually addresses the question (and there is more about this key skill in our page on Understanding Design Thinking).

Designers can also experiment faster with different options, to cycle through more ideas for clients in a shorter time. This means that they can work with more clients, and do more projects, ultimately making life more interesting.

There is of course a downside. Designers have lost work to AI, just like writers have. They may have had to reinvent themselves as ‘AI interpreters’, and changed how they work.

However, on the whole, the benefits of improved accessibility and better data analysis definitely outweigh the downsides.

Aspects to Consider When Using AI to Create Visual Materials

There are some important ethical and practical considerations to consider when using AI to generate or edit visual materials.

These include:

• The need to be transparent about your use of AI

  If you plan to share or publish your visual materials, you may need to declare the use of AI.
  Many publishers require this, especially if you are using an AI-generated graphic to illustrate an academic article. It is therefore important to be clear about the provenance of all the images you use, and understand the tools that you use to generate them.

• The importance of fact-checking everything produced by AI—including images

  Most people now understand that AI can hallucinate (a polite way of saying ‘make things up’).
  AI models don’t ‘know’ anything. They merely put information together based on things that they have ‘seen’ before. This means that what they produce may not be factually correct. You can try to control this with guardrails and by tailoring your prompts. However, ultimately you need to be prepared to fact-check anything that is produced by AI to ensure that there are no errors in it.

• The copyright and originality questions raised by AI

  There are some interesting copyright questions raised by AI.
  The first is whether an AI-generated image can be copyrighted at all. In the USA, the law is clear: an image purely generated by AI cannot be copyrighted. However, if you make extensive creative alterations to an image originally generated by AI, you may be able to copyright it. You therefore need to be very careful when using AI-generated images, to ensure that your intellectual property is protected.

  The second aspect of AI-generated images is that they often draw on pictures that the model has ‘seen’ before.
  You may therefore be infringing someone else’s copyright if you use an image straight from AI, without checking to see if anyone has already created something similar.

• The potential biases shown by AI

  All AI tools have potential to show bias.
  They reflect biases in their training data—and that often means that they will reflect their developers’ biases.

  Developers are (usually) aware of this issue, but cannot always address it adequately. Some examples of the way that AI models have surfaced hidden biases in the past include the voice-activated model that would not recognise women’s voices, the recruitment tool that only selected men for interview as software engineers (because most software engineers at the time were men), and the facial recognition software that only distinguished between the faces of white people.

  When you are using AI tools, you therefore need to be aware of whether the images show potential bias. It is a good idea to consider how you might overcome those biases through prompts.

• The higher energy requirements of using AI

  Using AI may seem effortless to you—and it often means that you take much less time to create something. However, there is a hidden cost to the planet. AI models use a LOT of energy to run. Every query put through an AI engine has a cost to it in terms of electricity and power use.

  It is therefore important that you don’t use them for trivial things.

---

A Final Thought

Using AI for visual materials has now become a standard part of the creative process in most organisations. It enables designers and creators to concentrate on the design strategy without worrying about the ‘grunt work’.

However, it is wise to be aware of both the benefits and the downsides.

In particular, using AI may seem effortless, but it has a hidden cost to the planet. We all need to be mindful of this, even as AI makes our lives easier.

---