Google has announced a new feature for its search engine, allowing users in the United States to create images from text descriptions. This generative AI feature, called Search Generative Engine (SGE), was unveiled at the I/O 2023 event and is set to expand to other countries soon. Users can input textual prompts like “hands holding flowers with a view of mountains in the background,” and Google’s search engine will generate up to four images that match the description. To ensure that users don’t mistake these images for human-generated content, Google adds metadata labels and watermarks to indicate that they were created by AI.
In August, Google introduced SynthID, an AI watermarking technology developed by Google DeepMind, designed to watermark and recognize AI-generated images. This technology is available exclusively to users of Google’s text-to-image AI generator ‘Imagen.’
To prevent misuse, Google has implemented safeguards and policies to block the creation of images that violate their guidelines, including those containing harmful or misleading content.
To enhance image credibility and combat misinformation, Google will introduce a tool called ‘About this image.’ This tool will provide information about an image’s history, such as when it was indexed by Google, where it was first seen, and where else it has appeared online, including on news, social, or fact-checking websites.
Additionally, SGE has been integrated with AI capabilities for generating written content in response to user queries. Users can request drafts for research articles, emails, or essays. For instance, if a user is dissatisfied with a product purchased online and wishes to send a return request, they can input a request like, “Write a draft to notify an online shopping company that the product they sent is inferior and ask the procedure to return it.” SGE can then synthesize information from various sources to generate a concise draft, saving users time and effort.
This new feature is similar to another generative AI feature introduced by Google in May, which provides additional context in search results.
In summary, Google’s latest announcement introduces a generative AI feature that allows users to generate images from text descriptions. This feature includes safeguards to prevent misuse and is part of Google’s broader effort to enhance image credibility and provide valuable AI-generated content to users.
, was unveiled at the I/O 2023 event and is set to expand to other countries soon. Users can input textual prompts like “hands holding […]&media=https://wonderfulengineering.com/wp-content/uploads/2023/10/ezgif.com-resize-99.jpg){kind=link}