There are numerous artificial intelligence image generators, each offering distinct features and varying levels of quality. One widely known option is available within Google Search, which excels in text placement on images and is currently an opt-in labs feature.
MidJourney v6 now includes the capability to overlay readable text on images, while the latest version of OpenAI’s DALL-E 3, accessible via ChatGPT, introduced this feature earlier in the year. However, in a quick comparison, neither quite matches the consistent accuracy of Google Search’s AI image functionality.
Introduced in October of this year and recently updated, this feature allows users to input specific prompts into the Google Search interface and swiftly receive relevant results. For instance, generating an image for a coffee shop named Bean and Gone with the tagline “So fine you’ll keep!” is one of the prompts available.
To access this feature, users need to visit the Google Search Labs webpage and engage with the conceptual AI search capabilities, which extend beyond mere image search to provide relevant information based on search queries.
Performance of Google Search AI Image Generator:
Initially, my expectations were modest due to its user-friendly interface. Previous attempts at adding text to images had been largely unsuccessful.
While DALL-E 3 performs adequately, it often requires multiple prompts and struggles to achieve precise results, especially when the overall image quality is high. MidJourney v6, on the other hand, prioritizes product imagery over text placement, making it slightly challenging to use for this purpose.
Upon testing Google Search AI with various queries, it consistently produced logical images with understandable text. While the majority of results were impressive, occasional issues like random letters, images devoid of text, or nonsensical phrases were observed.
Review of Google Search AI Image Generator:
Other AI image generators such as Starry and Ideogram specialize in creating images with readable text and excel in this aspect. Inform is particularly adept at crafting logos and ensuring accurate text placement.
Furthermore, companies like Nvidia offer proprietary designs, with DALL-E 3, MidJourney, and Google’s Imagen 2 model-based search tool being the primary options from major players currently available.
The standout feature of Google Search AI’s image engine is its user-friendly interface, seamlessly integrated within the search results. Users can simply click the ‘generate’ button when searching for specific images.
While the current implementation may not be groundbreaking, its foundation on Google’s Imagen technology hints at potential future integration into other Google-licensed apps and tools, potentially including Workspace in Presentations and Android platforms.
