Artificial Intelligence (AI) technology is evolving rapidly. Digital artists, creative professionals and academics are at the forefront of exciting advancements.
Adobe Photoshop’s Beta Generative AI is a tool with immense potential for creating captivating images for social media. This blog post, will explore how to use this AI tool, and discuss its benefits and limitations. Also, it will delve into the vital conversation about biases in AI.
Creating a Captivating Image: Using Generative AI
To start the creative process in Photoshop Beta, begin with a blank canvas and select everything. Use the Generative Fill function in the Edit menu to prompt the AI to generate alternative images based on your input. For this exercise my prompt was
“Post Apocalyptic Cyberpunk black female in her mid-twenties looking towards the camera in a head and shoulders portrait.“
A prompt will generate three alternative images, allowing you to choose the one that aligns best with your vision.
Fine-Tuning the Image: Enhancements for Perfection
After selecting an image, arising issues need addressing. In my chosen generated image, several areas required refinement. The AI generated eyes were aesthetically poor. When using the AI to replace the eyes Photoshop produced blue eyes instead of the desired aesthetics. I used the prompt with the AI of “Make the eyes match better and slightly wider.“.
However, this was probably not specific enough. I manually recoloured the eyes as the shape was adequate and it was quicker than creating a new prompt.
The next challenge was the nose, which was out of shape. I used the prompt with the AI to “Create a narrower nose.” The result was a significant improvement.
However, I used traditional retouching techniques in Photoshop to complete the image.
Liquify, allowed for general facial reshaping and I used colour grading with dodging and burning for the finishing touches.
Efficiency in Image Creation: The Time Factor
With Adobe Photoshop’s Beta Generative AI, creating a final image for social media or video clips becomes an efficient process. Despite some imperfections, investing approximately 20 minutes I was able to direct the AI to craft a visually striking image and demonstrate the tool’s potential. Moreover, this image serves as a starting point for further iterations within Photoshop or other Generative AI tools, offering endless possibilities for artists to explore.
Challenges and the Road Ahead: Addressing Biases in Generative AI
While Generative AI presents exciting opportunities, it also brings challenges. One significant concern revolves around racial biases in the AI’s image generation process. In this case, the AI generated a cyberbody that didn’t match the desired representation of a black female subject. Such biases are reminiscent of everyday challenges for people of colour. For example, buying items like plasters that primarily cater to white skin tones. Removing the assumption that “white is right” is crucial in mitigating these biases.
Furthermore, limitations on prompts and censorship pose challenges for artists using AI tools. In an earlier experiment, I discovered that the word “Zombie,” was prohibited by Adobe. Censorship can hinder artistic and creative freedom. Deciding which words are restricted and determining the ethical and legal boundaries surrounding AI usage requires careful consideration. We need open dialogue is necessary to foster collaboration between AI developers and the creative and business communities to effectively address these concerns.
Generative AI and the Transformation of Creative Industries
The impact of AI extends beyond image production, signalling a transformative shift in creative industries. Generative AI enables artists to create social media ready images as well as other assets, questions arise regarding the need for models, photographers, and associated teams. Integrating AI voiceovers and video production further augments the potential for individuals to craft their own virtual social media stars and influencers. As these possibilities unfold, it’s crucial to engage in discussions and proactively address the issues.
—
More on bias in technology
Leave a Reply