The capabilities of artificial intelligence (AI) image generation tools have grown at a terrifying pace. These are some of the best on the market.
If 2023 has a definitive buzz phrase, it has to be “generative artificial intelligence”.
The explosion of interest in OpenAI’s chatbot ChatGPT set the stage for a year in which millions of people started using generative AI tools for the first time.
While entire articles or exam papers can be generated by a simple text prompt, text prompts can also be used to generate images. Many platforms have emerged that allow users to create artwork, posters, logos, presentations, and photorealistic images by feeding them with just a few words.
They can do this thanks to machine learning. Neural networks are trained on massive datasets, learning to identify patterns and features in images. The algorithm then uses this learned knowledge to predict what a text prompt is looking for.
We’re taking a look at some of the most popular image-generation tools available today. Before diving in and using the platforms, it’s worth paying attention to the copyright situation around AI-generated images. There are already a number of lawsuits ongoing, with artists claiming damages from companies whose AI platforms they say were trained on their copyrighted material.
Let’s take a look at five of the best image generators available today.
Prompt: “A power mage casting a spell”
One of the first AI image generators to go big, Midjourney attracted millions of users to its platform as users were blown away by its capabilities, and enticed by its free trial.
That free trial no longer exists, so you’ll have to pay for a subscription now to try it out for yourself, with prices starting at $8 (€9.33) a month.
The process can be fiddly, as it functions through Discord rather than a standalone platform. With some trial and error, or by taking lessons from the numerous prompting guides available online, you will quickly see how powerful Midjourney is, creating photorealistic images, art, cartoons, or sketches at your will.
No wonder it already has more than 15 million users.
Prompt: “Painting of a plane taking off at dusk in the style of Vincent van Gogh”.
From the company behind the wildly popular ChatGPT, OpenAI’s DALL·E 2 is another popular image creation tool.
Users can generate a number of images on the free plan and will need to sign up for a paid plan to do so in bulk. The platform boasts millions of users, although it doesn’t seem to work as well as Midjourney for photorealistic-style images. What it does excel at is artwork, with a knack for copying the style of famous painters.
And just as OpenAI’s GPT-4 powers the Microsoft Bing search engine’s AI chatbot, DALL·E 2 also powers Bing’s image generation tool.
DreamStudio (Stable Diffusion)
Prompt: “A steaming mug of tea on a rustic wooden table. Trees are visible through a window in the background”.
Another of the early big hitters, Stable Diffusion is a popular image generation model, with a free tool on the web browser. DreamStudio uses the Stable Diffusion model and has a host of options and a professional user interface.
When you sign up you get a certain amount of free credits, and you’ll then need to pay to top them up. The results are impressive, especially when generating human faces – although like all these image generators, it seems to have a particular problem with human hands.
Dream by WOMBO
Prompt: “Alien space station”
Geared towards art generation, like the other platforms, you just stick in a simple text prompt and it will deliver images dreamed up by the algorithms. As the name suggests, the images it produces are somewhat dreamlike, with abstract and often creepy results.
With dozens of art styles to choose from (including things like “bad trip” or “steampunk”) it’s a veritable playground of art creation. It’s available as a web app and a mobile app.
Prompt: “Man driving a supercar down a road with the Eiffel Tower in the background.”
The popular graphic design platform that is largely used to create social media posts and presentations has jumped into generative AI with both feet.
We’re including this one as a special mention because its capabilities in generating the aforementioned media types are truly impressive. It also has a free image creation AI which is perhaps a little less refined right now.
It has quite strict rules about the types of images it will create.
For example, it refused to depict Emmanuel Macron driving a Ferrari down the Champs-Élysées, so we had to settle for a “man driving a supercar down a road with the Eiffel Tower in the background” instead.
It looks like it is still in the work-in-progress phase.
Source: Euro News