GenAI "Copilot"
Hellow everyone. I am the editor-in-chief of "KemonoGirls".
Recently, many people may have heard the terms "AI image generation" or "GenAI" or seen AI images on social media. Some of you may want to create AI images but think it is difficult, have looked online but don't know where to start, or are worried about the cost.
I recommend "Copilot" for those people, which allows you to enjoy image generation for free. You can use it as many times as you want without ads, so it's a great first step to get used to image generation.
Copilot AI
Copilot is a Microsoft AI that has AI chat and image generation functions. Anyone with a Microsoft account can use it. It comes standard with Windows 10 and Windows 11.
How to use
Search for "Copilot" from the Windows Start button
Sign in to Microsoft using the Edge browser
For Google, Safari, etc., sign in to Microsoft on the Bing
Sign in to Microsoft in Copilot applications
Let's try it out
In image generation, AI extracts elements from the text you enter and automatically creates an image. This text is called a "prompt".
There is a limit to the number of times you can enter a prompt, 5 times in the browser and 30 times in the app. When you reach the limit, you can start a new topic, but you will have to start over from the beginning.
Typically prompts have a set format, but the great thing about Copilot is that it lets you write as if you were having a normal conversation.
Now let's actually enter some prompts and generate some images.
Fox illustration
For example, if you write the prompt
"Illustration of an orange fur fox sitting on sofa in living room"
an image like this will be generated.
The number of images generated at one time is 1 to 4, and this number is automatically determined, so you cannot specify the number. You can also edit the prompt you wrote or enter the elements you want to add to generate new images based on the images you created.
If you want to make this illustration look like a realistic photograph, just enter "Realistic photograph of an orange furry fox sitting on a sofa in the living room" or "make it a realistic photograph" and it will be generated like this.
This image was generated by taking the elements of a fox, a living room, and a sofa. You can also add details such as color, shape, and facial expression to these elements. For example, it looks like this.
sitting on a sofa in a living room
What do you think? The fox's eyes, which were brown at first, have turned blue, the sunlight coming through the window has become stronger, and the sofa has turned blue. The fox's face, which was expressionless until now, has also changed to a cute, sleepy expression.
A concise, yet detailed prompt is more likely to produce a result that matches your imagination. If the prompt is too long, it will still produce a result, but it will prioritize elements written earlier in the prompt. If you want to draw a character, write the character's characteristics first, then write the background details.
Images can be saved
There is an icon in the upper right corner of each generated image. Tap this icon to save it. The image will be saved in Bing's "Collection". There is no limit to the number of images you can save, so you can save as many as you like.
Of course, you can save the images individually to your desktop, but you can also use an app on your smartphone to generate images during the day and save them all to your desktop at night.
Reimage
In addition to text input, you can also add images to generate images. The added image is analyzed by AI, and elements are extracted and generated. By describing the prompt, you can generate it more accurately. In this case, the prompt will be the description of the image to be added.
Important point
Here are some points to note when using the ai. There are some images that cannot be created and some things that cannot be done with the free version.
Wi-Fi environment recommended
No violent or sexual images are generated
What you can do with the paid version
It's unclear whether commercial use is permitted
Copyrighted characters can be generated
1. Wi-Fi environment recommended
Not only Copilot, but when using image AI on a smartphone, a lot of communication is used for processing. It drains the battery. Also, since it is crowded at night, restrictions are likely to be imposed and generation will stop. When that happens, it will not resume for 4 to 5 hours. If the restriction is imposed in the middle of the night, give up and go to sleep. It should resume in the morning.
2. No violent or sexual images are generated
Copilot cannot generate violent or sexual imagery. R18 rated content is prohibited. It cannot generate direct prompts, innuendos or paraphrases. For example, for a female body, you can specify height but not bust size or body type. It also has difficulty generating revealing clothing.
3. What you can do with the paid version
In the paid version, you can specify the aspect ratio when generating images and you can also edit the generated images, but in the free version, the aspect ratio of generated images is limited to square and cannot be edited within Copilot.
4. It's unclear whether commercial use is permitted
While other AIs state in their terms of service that commercial use is permitted, Copilot's terms of service state that "use of the online services does not give you any ownership rights to intellectual property or other data." Maybe non-commercial use is better.
Quote source: Copilot AI Experiences Terms
5. Copyrighted characters can be generated
Famous characters such as Super Mario and Pikachu will be generated. I tried it myself, but it was not generated if it was too niche. Please be careful when using it as stated in the terms of use.
What did you think? This time, we introduced the overview of Copilot and examples of its use. Even if you are a beginner to AI, I think it is a good idea to start by learning how to write prompts in Copilot and understanding how it works.
Copilot is also an AI chat that can give you accurate answers to your search terms using online information. You can also ask Copilot directly about the prompts, so feel free to ask for tips on writing.
I wrote earlier that "clothes with a lot of exposure are less likely to be generated," but that doesn't mean they won't be generated at all. In fact, there is a way to make them more likely to be generated by tweaking the prompts. I'd like to introduce that in another article.
Thank you for watching until the end ♪
この記事が気に入ったらサポートをしてみませんか？