What is Bing Image Creator & How Do I Use It?
Bing Image Creator is a free AI art generator developed by Microsoft Bing powered by DALL-E 3. You have to pay for ChatGPT Plus or Enterprise for access to DALL-E 3 (or earlier versions) on openai.com, but it’s completely free via Bing Image Creator. So, if you’re looking for a free way to create AI generated images and artwork – this is the article for you!
How do I use Bing Image Creator?
Using Bing Image Creator can be surprisingly straightforward considering Bing’s history UX struggles (personal opinion). First, you have to be logged into a personal Outlook, Hotmail or Microsoft account (below is the error you’ll see if you’re not logged in).
Next, you can access Bing Image Creator’s AI art generator features directly through Bing Chat, Image Creator from Designer or the Microsoft Edge sidebar (screenshots below).
Prompting Bing Chat to create images is as easy as starting your prompt with the “show” or “create an image of…” For example, a prompt in Bing Chat for “show me what an alligator in a lambo looks like” and Bing was able to determine that an image was the best response. Using phrases like “show me” or “create an image of” aren’t necessary when prompting Bing directly via Image Creator from Designer.
Follow-up prompts after the initial image generation aren’t as smooth in Image Creator from Designer as they are in Bing Chat. For example, when using Bing Chat, you can just say “more it more colorful” while the Image Creator from Designer experience requires the full prompt each time.
Bing Chat limits users to five prompts per session for image creator the same way they do for text generation. Bing almost always provides four images per prompt based on my experience; however, I have seen instances where only one to three images are generated.
Does Bing save Image Creator images?
The nice part about Microsoft requiring users to sign-in to use Bing Image Creator is that they save your recent creations. Based on my testing, it looks like Bing holds onto the images from your last 20 prompts. This can be really helpful when you don’t have time (or disk space) to download all the images generated, but also don’t want to lose them.
What size are DALL-E 3 images?
Images created using Bing Image Creator have the same size and aspect ratio limitations as images creating directly with DALL-E 3.
Bing Image Creator images are 1,024 x 1,024 pixels with a 1:1 aspect ratio. Although DALL-E API supports images that are 512×512 and 256×256 pixels, however it doesn’t appear these are used by Bing.
I recommend checking out Midjourney which offers custom aspect ratios if you’re looking to create AI-generated art that isn’t square (i.e., 1:1 aspect ratio).
What are boosts for Bing Image Creator?
Boosts are single-use tokens provided to users for faster image generation processing times. Users will receive boosts when they sign up and will continue to receive additional boosts each week. Microsoft Edge documentation for boosts states that users begin with 25 boosts and will continue to receive up to 15 boosts weekly without exceeding a total of 15.
From my experience it seems like the image generation is quicker in Bing Chat as opposed to Image Creator from Designer, but I’ve yet to do extensive testing to proof it out. Images generated through Bing Chat do use boost tokens though even though it’s not mentioned.
Users (on Reddit & SEO forums) that have run out of boosts have claimed the processing time can vary wildly based on usage – ranging anywhere from 30 seconds to 4 hours.
10 examples of Bing Image Creator prompts & responses
Here’s some examples of images created using Bing Image Creator (powered by DALL-E 3) and the prompts used for each:
“Show me what an alligator in a lambo looks like” prompt
“Countdown clock” prompt
“Colorful countdown clock” prompt
“Countdown clock without words that is colorful and shows the time 12:00” prompt
“Search engine optimization” prompt
“What does SEO looks like to non-technical people” prompt
“What SEO people look like” prompt
“Image creator” prompt
“A graphic designer” prompt
“Penguins playing polo” prompt
How Nuaveu uses AI-generated images
I primarily use AI-generated images for Nuaveu blog posts and decorative images. Being able to create high quality images, quickly and for free has been a game-changer. I always encourage small business owners operating on a shoestring budget to consider the benefits of AI-generated art. I’ve also found AI-generated images to be really helpful when brainstorming logo ideas and changes to existing logos. With that being said, I also firmly believe that graphic designers are critical to our industry. Even though I’ve enjoy creating and benefit from using AI-generated art I’m painfully aware of the harm it’s causing. Hire a human over using AI whenever possible – the quality of work will always be better.
Using Bing Image Creator for Marketing
When Bing Image Creator first launched the terms of use stated that images were only for non-commercial use. However sometime last year this changed and neither the content policy nor the Copilot AI Terms mentions this restriction anymore. Also, from my experience Image Creator doesn’t include watermarks in exports. Even when a watermark shows up in the image preview it isn’t present when downloaded. Thankfully this means you can now use Bing Image Creator for marketing. So, if you need a graphic for your email newsletter, social media post or blog article – you should it a try. It’s a great free alternative to using DALL-E 3 via ChatGPT Plus (currently starting at $20/month).
Please note that this is just my interpretation of Microsoft’s policies and should not be taken as legal advice.
Issues with generative AI images
AI-generated images have their flaws even aside from the moral or ethical issues briefly mentioned above. Here’s some of the top issues I’ve noticed (many of the images referenced are included in the examples above):
- Words are often illegible, overlapping, missing letters, contain typos etc.
- E.g., “SO” instead of “SEO”, “Sar Wars” instead of “Star Wars” and in many cases there’s just scribbles instead of words
- Lots of duplication/repeated elements
- E.g., four Mona Lisas appearing in a single image of an “image creator”
- Objects that should have symmetry don’t and vice versa
- E.g., keyboard keys not lining up or being straight and unusual/unrealistic appendages on people/animals
- Stereotyped images
- E.g., 3 of 4 images for “graphic designer” were of a young(ish) white guys with a goatee
- Inconsistent/random themes within a single image
- E.g., putting multiple Mona Lisa and Star Wars elements together in a single image without being asked to
- This isn’t always a negative and can be genuinely hilarious, but something to be aware of
How to use DALL-E 3 for free
If you’re looking to use DALL-E 3 for free simply log into your Microsoft Outlook account and go to https://www.bing.com/images/create and submit a prompt. I hope you found this article entertaining and helpful. Best of luck with your generative AI adventures and happy prompting!
Update 1/25/24: added information about watermark and non-commercial use being removed from terms.