Media Creation with Generative AI

Below are a few of the popular generative AI tools. These tools have a range of functionalities, but all start with the creation and refinement of images, which starts with a prompt. Additionally, applications like Adobe's FireFly and Canvas Magic Studio will allow users to create and remove backgrounds and distracting elements, add neural filters, and expand a scene using generative fill. 

• Adobe Firefly is a stand-alone website and a suite of generative AI tools developed by Adobe, designed to integrate seamlessly into Creative Cloud applications. It allows users to generate creative content, such as images, text effects, and designs. In addition to Firefly, many of the generative AI functionalities can be accessed through standard Adobe applications, such as Adobe Photoshop, Adobe InDesign, Adobe Illustrator, and Adobe Premiere Pro. 
• Gemini by Google is an AI-powered assistant that can help with a variety of tasks, including image generation. These models integrate text, images, audio, and video understanding to enable more intuitive interactions and versatile capabilities within the Google workspace. 
• Canva is a graphic design platform that has always focused on making it easy for anyone to create designs for anything from social media posts to infographics to presentations. Canva's AI tools are conveniently grouped within the software's "Magic Studio" section, which is displayed in Canva's menu. It allows users to generate creative content, such as images, text effects, and designs.
• Midjourney AI is an independent research lab and creative platform that generates high-quality, AI-powered images from text prompts.
• CoPilot by Microsoft is a tool in Microsoft that uses AI to create images based on text prompts. Copilot is available in PowerPoint and Word to create images for documents and presentations, but it is also a standalone website. 

Creating text-to-image prompts involves crafting detailed and precise descriptions that convey your creative vision to the AI model. A good prompt balances clarity and creativity, ensuring the output aligns with your expectations. Patience is a virtue. Prompts take time to hone and get the exact image you are trying to create. Very few times will you get the image or design you want from the first prompt entered.

Below are some basic concepts to think about as you craft your prompts. Start with a general idea, then iterate and refine based on results. 

A comprehensive prompt typically consists of:

• Main Subject: What or who is the focus? (e.g., "a bulldog sitting on in a chair").
• Attributes: Details about the subject's appearance, behavior, or emotions (e.g., "a bulldog, looking inquisitive").
• Setting: The environment or background (e.g., "in a sunny room with potted plants").
• Art Style or Medium: Specify the visual style (e.g., "watercolor painting," "realistic digital art," "pop art," or "cyberpunk").
• Lighting and Mood: Describe the atmosphere (e.g., "soft, warm lighting," "dramatic lighting," or "cold lighting" ).
• Additional Elements: Any specific objects, actions, or themes to include (e.g., "a cup of tea on the table beside it").

Many of the tools mentioned above have landing pages where you can see the results of other users' prompts. Depending on what tool you are using, try copying, but changing one of the above elements can drastically change your results. 

Image-generating AI presents several ethical issues that entitle careful consideration, including but not limited to:

• Copyright Infringement: AI models often train on vast datasets that include copyrighted works without explicit permission from the original creators. This raises questions about whether using these works is fair use or theft.
• Misinformation and Deepfakes: AI-generated images, audio, and, video can be used to create deceptive news, deepfakes, or misleading media that spread misinformation.
• Bias and Stereotyping: AI models reflect the biases in their training data, which is majority western based, and potentially perpetuating harmful stereotypes or excluding underrepresented groups.
• Sensitive Content: AI might generate inappropriate or exploitative content, including sexualized depictions of individuals without their consent.
• Attribution and Accountability: Creators of AI images and designs should acknowledge the technology and the developers behind their AI-created content. This ensures transparency, helping viewers understand the origins of the content they see. In addition, failure to properly cite AI-generated images can lead to legal issues, especially if the images are used commercially. 

Uploading images, audio, and video to the Content Credentials site will reveal helpful information that's usually hidden. Critical information about the content you see online is often inaccessible or inaccurate.

• Output thumbnail: A visual thumbnail of the output that display.
• Issuer: The organization responsible for issuing the Content Credential. Ie Adobe Inc., Microsoft, etc. 
• Content summary: A notice that generative AI was used to create the asset.
• App or device used: The Adobe software application or hardware device used to generate the asset.
• AI tool used: Adobe Firefly, Microsoft CoPilot, etc. 
• Actions: The general editing and processing actions taken to produce the asset. 
• Creation Date and Time