The technology sector is currently witnessing a competitive race in the field of artificial intelligence, with industry leaders like Google, Meta, Microsoft, and others continuously introducing new features and models.
In a recent development, Google has unveiled an intriguing feature in its Search Generative Experience (SGE) that allows users to generate images directly from a text prompt. This feature bears a resemblance to the Bing Image Creator that Microsoft introduced a few months ago.
Google’s official blog post shares its vision, stating, “As we continue to explore the integration of generative AI capabilities into Search, we’re exploring new avenues to enhance user experiences, such as generating images to bring ideas to life or assisting with written drafts as a starting point.”
In essence, Google is simplifying tasks for users through its AI-powered Search Engine (SGE), or Search Generative Experience. SGE represents Google’s innovative search engine that leverages AI to provide more comprehensive and informative responses to user queries. The latest update to SGE further enhances its interactivity and utility.
Google recently introduced AI-backed summaries of search results through SGE, offering users concise overviews of their search queries.
Additionally, it enables users to check grammar and other language errors in their search query sentences. With the latest update, users can now issue requests using phrases like “draw” or “create a [x],” prompting SGE to generate entirely new images.
The image generation within SGE is powered by Google’s Imagen family of models, similar to features found in Google Slides and Meet.
How Image Generation on Google Search Works
Google’s generative AI in search is facilitated by its Imagen text-to-image diffusion models, granting users access to the SGE. This allows users to request images through text prompts in search queries, much like how Bing Image search operates. For instance, if a user desires an image of a dog dancing on a beach, they can simply type their request into Google Search. Upon issuing the prompt, Google will give the user four different image options. Furthermore, users can modify their descriptions to tailor the image to their preferences.
To experiment with AI image generation in Google’s SGE, users can input “draw” or “sketch” in the Google Search box. After sending the prompt, Google will display up to four images simultaneously, appearing above the standard search results with a distinct background colour.
Additionally, while users are browsing image search results, Google may suggest AI-generated images.
It’s worth noting that there might be a delay in SGE updates within image searches, requiring users to exercise patience. Nevertheless, this feature will be invaluable to users seeking specific images matching their precise requirements.
Users can tap on any of the generated images to observe how generative AI has expanded on their initial query, providing descriptive details. They can further edit the description to add more specifics and bring their vision to life.
It’s important to highlight that Google is deploying these new AI features responsibly, ensuring that the image-generation tool adheres to the company’s prohibited use policy for generative AI, avoiding any generation of images that violate ethical guidelines.