5 Workflows to Create AI Images That Don't Look Generic

18 Apr 202455:16

TLDRIn this engaging video, the presenter discusses strategies for creating unique and non-generic AI-generated images. They delve into the process of 'monster creation', emphasizing the importance of originality when designing creatures and characters. The session includes practical demonstrations, where the presenter uses various techniques such as image-to-image transformations and control nets to guide the AI in producing distinctive artwork. They also touch upon the challenges of deviating from the normative expectations of AI models and highlight the significance of creative decision-making in the art creation process. The video concludes with a sneak peek at an upcoming tool for regional control, which promises to offer artists an unprecedented level of precision in guiding AI image generation.


  • 🎨 **Embrace Non-Generic Creation**: The session focuses on creating AI images that stand out from generic outputs by infusing unique concepts and making creative decisions.
  • 🚀 **Start with Crazy Ideas**: For non-generic creations, begin with wild ideas, especially in monster and creature design, to push the boundaries of what exists.
  • 📝 **Use Prompts Effectively**: Utilize text prompts to guide the AI in creating images, but also to push it towards less conventional outcomes.
  • 🌟 **Iterate and Refine**: The process involves multiple iterations, using control nets and image-to-image transformations to refine and detail the creations.
  • 🧩 **Combine Elements**: Mix and match different elements, like animal characteristics, to create unique creatures that don't follow standard norms.
  • 💡 **Inspire with Examples**: Using example images as control nets can guide the AI to generate more specific and desired outcomes.
  • 🔍 **Focus on Details**: Pay attention to the details, such as the placement and structure of features, to achieve a more refined and non-generic result.
  • 🧬 **Manipulate the AI's Understanding**: By providing examples and training, you can alter the AI's understanding of what is normal, allowing it to create more unconventional images.
  • 📈 **Experiment with Regional Control**: Upcoming tools like regional prompting will offer more control over specific parts of the generated image, leading to even more unique creations.
  • 🎭 **Artistic Influence**: The role of the artist is to make decisions at every step, guiding the AI to create something that is less normative and more inspired.
  • ✂️ **Post-Processing**: Many of the generated images can benefit from post-processing in programs like Photoshop to finalize the unique creations.

Q & A

  • What is the main focus of today's event?

    -The main focus of today's event is to explore ways to create AI images that are non-generic, particularly through the process of monster creation.

  • How does the speaker suggest using an initial image to guide the AI in creating unique creatures?

    -The speaker suggests using an initial image as a base to guide the generic structure of the creature while infusing different concepts and using prompts to add unique aspects or character traits.

  • What is the significance of making creative decisions in the process of creating art, as mentioned by the speaker?

    -Making creative decisions is significant because it allows the artist to differentiate their work from what's expected or normative. It's about guiding the creative process and injecting personal influence into the work.

  • How does the speaker use the term 'non-normative' in the context of AI art creation?

    -The term 'non-normative' is used to describe the intention to create AI art that is different from the average or expected outcomes, pushing the boundaries of what the AI model typically generates.

  • What is the role of the 'control net' in refining the AI-generated creature?

    -The 'control net' is used to maintain structural consistency and guide the AI in refining the creature's shape, ensuring that details like limbs and features are consistent and anatomically correct.

  • How does the speaker demonstrate the iterative process of AI art creation?

    -The speaker demonstrates the iterative process by taking an initial AI-generated image, making modifications, and then using those changes as a base for further refinement. This process involves multiple cycles of generation, evaluation, and manual adjustment.

  • What is the purpose of using an 'IP adapter' in the AI art creation process?

    -The 'IP adapter' is used to influence the AI's generation process by providing a reference image that the AI can use to incorporate specific visual elements or themes into the new creation.

  • How does the speaker address the challenge of creating non-anatomical human characters?

    -The speaker addresses this challenge by manually adding extra limbs to a generated human figure and using the AI's ability to adapt and generate based on provided examples, suggesting that training the AI with more examples of non-anatomical features can improve results.

  • What is the 'Regional prompting' tool mentioned by the speaker, and how does it offer more control over AI art creation?

    -The 'Regional prompting' tool is a feature that allows for more precise control over different parts of the AI-generated image. It works by applying different prompts to different regions of the canvas, enabling the creation of complex and unique compositions.

  • Why is it important to train the AI model with various examples when aiming for non-generic results?

    -Training the AI model with various examples is important because it expands the model's understanding of what is possible and normal within its universe. This helps the AI to generalize and create more diverse and non-standard outcomes.

  • How does the speaker suggest using the AI's limitations to inspire creativity?

    -The speaker suggests using the AI's limitations as a creative challenge, pushing the AI to generate images that it might not typically produce. By doing so, the artist can explore new and unique creative spaces.



🎨 Exploring Non-Generic Monster Creation

The speaker discusses the importance of creating non-generic content, using monster creation as an example. They emphasize the need to avoid being identical to everything else and share their process of generating unique monster ideas. The speaker demonstrates using an image-to-image transformation technique to create a 'Tiger Raptor' creature, suggesting ways to adjust and refine the creature's features by controlling the prompts and adapting the structure.


🤔 The Art of Creative Decision Making

This paragraph delves into the concept that art, and by extension entrepreneurship, is fundamentally about making creative decisions. The speaker highlights the challenge of being non-normative while using AI art, given that AI models tend to produce generic outputs if not guided properly. They discuss the balance between pushing the boundaries in some aspects while maintaining structure in others, and how this process is integral to the creation of unique and compelling art or products.


🦎 Morphing a Creature with AI and Iterative Design

The speaker continues the discussion on AI art creation, focusing on the iterative process of transforming a basic creature shape into a more complex and interesting design. They mention the use of an 'armadillo' creature as a starting point and how they refine it by adding features like horns and cleaning up the design. The paragraph also touches on the idea of using AI as a tool to inspire and guide the creative process, rather than relying solely on its outputs.


🧞‍♂️ Training AI for Non-Generic Outputs

Here, the speaker addresses the challenge of creating non-generic art with AI by training the AI to understand non-standard concepts. They discuss the use of an image prompt to guide the AI towards creating a unique creature with transparent spikes. The speaker also talks about the potential of regional prompting as a future tool for gaining more control over AI-generated content, allowing for the creation of non-normative and unique pieces.


🖌️ Combining AI with Manual Artistic Touches

The paragraph discusses the hybrid approach of using AI for conceptual creation and then refining the results manually using tools like Photoshop. The speaker demonstrates how to take a creature generated by AI and further develop it by adding details and making adjustments to achieve a desired look. They emphasize the value of human creativity in guiding and finalizing the AI's outputs to make them truly unique.


🤹‍♂️ Experimenting with AI to Achieve Non-Generic Designs

The speaker shares their process of experimenting with AI to create non-generic creatures. They use various prompts and images to guide the AI in generating unique designs, such as a beast with an exoskeleton. The paragraph also explores the idea of regional prompting, where different parts of an image are controlled by separate prompts, allowing for greater creativity and control over the final output.


🧬 The Challenge of Non-Anatomical Human Features

In this part, the speaker tackles the challenge of creating non-anatomical human features, such as a character with four arms. They discuss the limitations of AI when it comes to generating non-standard anatomical structures and suggest training the AI with examples of the desired features to improve the results. The speaker also teases the upcoming regional control feature as a powerful tool for guiding AI in creating complex and non-generic imagery.


📈 Regional Control for Advanced AI Art Direction

The speaker provides a sneak peek at the upcoming regional control feature, which allows for granular control over different parts of an AI-generated image. They demonstrate how this feature can be used to combine elements from various prompts to create a unique creature with a hyena's head and an insectoid body. The paragraph concludes with the speaker expressing excitement about the potential of this tool for artists and creators.




The term 'non-generic' refers to something that is unique and not typical or standard. In the context of the video, it is used to describe the goal of creating AI-generated images that stand out from common or average results. The video emphasizes the importance of avoiding generic outcomes in AI art to foster creativity and originality, as illustrated by the speaker's efforts to create unique monster designs.

💡Monster creation

Monster creation is the process of designing and conceptualizing fictional creatures or monsters. It is a central theme in the video, where the speaker discusses generating non-generic monster ideas. The script mentions 'coming up with ideas for monsters, and creatures,' highlighting the imaginative aspect of this process.

💡Image to image

'image to image' is a technique used in AI image generation where an initial image is transformed into another image based on a given prompt. The video script describes using this technique to take an image of normal animals and transform it into monsters, demonstrating how it can be used to create unique and non-generic AI images.

💡Control net

A 'control net' is a tool used in AI image generation to guide the process by providing a structural base that the AI can use to create an image. In the video, the speaker uses a control net to maintain the basic shape of a creature while infusing different concepts into the design, such as adding horns to a tiger-like monster.

💡Creative decision making

Creative decision making involves making choices that determine the direction and nature of a creative work. The video emphasizes that art, including AI-generated art, is fundamentally about making creative decisions. The speaker discusses how each choice can make the artwork more or less like existing works, and striving for non-generic results requires deliberate and thoughtful decision making.

💡Denoising strength

Denoising strength is a parameter in AI image generation that controls the level of noise reduction in the generated image. A higher denoising strength typically results in a cleaner, more refined image, while lower settings allow for more variability and 'noise.' In the video, the speaker adjusts the denoising strength to achieve the desired level of detail and control in the AI-generated creatures.

💡IP adapter

An 'IP adapter' in the context of the video refers to using an existing image as a reference or guide for the AI to generate a new, similar image. The speaker uses an image of a hagfish as an IP adapter to influence the look of a creature, demonstrating how this technique can be used to inject specific visual elements into AI-generated images.

💡Regional prompting

Regional prompting is an advanced technique mentioned at the end of the video that allows for control over specific areas or 'regions' of an AI-generated image. This tool is said to provide the utmost degree of control and is expected to help creators achieve non-generic results by influencing specific parts of the image independently.

💡AI training

AI training involves providing the AI with examples or data so it can learn and improve its performance. In the video, the speaker discusses the concept of training the AI to understand non-standard concepts, such as a four-armed human, by showing it examples of such entities. This helps the AI to generalize and create more varied and non-generic outputs.

💡Concept painting

Concept painting is a form of visual art used to create a representation of an idea or concept, often for the purpose of planning or illustrating a design. The video script mentions 'concept painting' in the context of using a clean, white background to force the AI into a particular direction of image generation.

💡Deep sea creature

A 'deep sea creature' is a marine animal that lives in the deep parts of the ocean. In the video, the speaker jokingly refers to one of the AI-generated creatures as a 'deep sea fake,' highlighting the unusual and imaginative nature of the creature, which has transparent spikes and a unique appearance.


The event focuses on creating non-generic AI images, particularly through the concept of monster creation.

The importance of being non-generic is emphasized to avoid producing images that are similar to everything else.

A technique for generating ideas is demonstrated using a sheet of monsters as a starting point.

The use of an 'image to image' tool is shown to transform and create a larger, more detailed monster image.

A tiger Raptor creature is conceptualized by infusing different concepts using the prompt.

Adjusting prompt weights can control the shape and characteristics of the generated creature.

Adding horns and other features to the canvas can further detail the creature.

The process of making creative decisions is compared to an entrepreneur's approach to finding inspiration and uniqueness.

The concept of 'art is about making decisions' is discussed, highlighting the role of the creator in guiding the process.

Different ways to be non-normative with AI art are explored, including using an initial image to shape the noise.

The challenge of using text prompts to create non-generic AI art is acknowledged.

A method for generating a creature with four arms is attempted, showcasing the AI's struggle with non-anatomical human features.

The use of an IP adapter is demonstrated to influence the AI's generation process.

Regional prompting is introduced as an upcoming tool that provides more control over AI image generation.

A sneak peek at regional control nodes is given, showing their potential for creating non-standard creatures.

The session concludes with a discussion on training AI models to understand and generate non-generic concepts.