Columbia Technology Ventures

Opal: AI image generation from text prompts for enhanced news illustration

This technology is an AI text-to-image generating platform for news illustrations called Opal, which can enhance readership through the generation of captivating images.

Unmet Need: Streamlined text-to-image AI prompt-based image generator for journalism

Images and visual elements are crucial aspects to journalism and the creation of captivating news articles. Recent advancements in artificial intelligence technologies have allowed for tools which can generate text and images based on certain prompts and supplied criteria. While these tools can be powerful, they are hard to utilize, leaving them inaccessible by those who do not know how to use them or how to generate good prompts to create desired images. Although these systems could be applied to aspects of journalism, there is currently no single platform dedicated for text-to-image generation specifically geared towards news and media stories.

The Technology: User-friendly and robust text-to-image AI platform for journalists

This technology is a prompt-based program called Opal, which uses AI platforms to generate images for journalism applications. The system allows for the use to input the text of the article or media where images are desired, and the platform will identify different areas where images could be utilized based on emotional tone, subject matter, and topic transitions. It will then generate images to fit these goals based on the approval of the user. This allows for robust use of modern AI text-to-image generation technologies in the journalism industries, which may increase readership and popularity of media platforms.

Applications:

  • Image generation for news illustrations
  • Product advertising and marketing
  • Captioning of journalistic images and media
  • Portrait construction
  • Prototype imaging and design

Advantages:

  • User-friendly
  • Easy to use interface and platform
  • Robust algorithms for text-to-image generation based on entire article text
  • Uses article tone, emotion, keywords, and art styles to suggest images
  • Enhances query searching through prompt engineering

Lead Inventor:

Lydia Chilton, Ph.D.

Related Publications:

Tech Ventures Reference: