Google's SGE: From Text to Imagery
No attachments for this post
Google has unveiled the Search Generative Experience (SGE), which transforms text prompts into images. This development comes after Microsoft's similar initiative with Bing Chat, utilizing OpenAI's DALL-E model.
Here's a quick overview:
- Users who've joined the SGE via Google's Search Labs can enter a text in Google's search bar.
- In response, SGE, powered by Google's Imagen AI models, crafts several image options.
- An illustrative GIF by Google demonstrates creating an image from the prompt “draw an image of a capybara wearing a chef's hat and cooking breakfast.” Users can fine-tune these queries for more accurate visuals.
- This image generation can also be accessed on Google Images. While browsing, a prompt might appear, enabling users to generate images that will be displayed in a sidebar.
- Hema Budaraju from Google emphasized the company's commitment to responsible AI image generation. This tool is designed not to produce images that breach Google's generative AI policy.
- All AI-created images will carry metadata and watermarks, signifying AI as their source. At its debut, the tool will not produce photorealistic faces or images of renowned figures. Additionally, the feature is restricted to users aged 18 and above.
- The SGE's capabilities also include generating text drafts based on queries. Users can define the length and tone, and if satisfied, transfer the draft to Google Docs or Gmail.
- Craig Ewer clarified that SGE utilizes a combination of LLMs, trained on public data and human-guided input.
Since its introduction in May, SGE has seen various enhancements. Despite its sluggishness, which some users find limiting, the recent additions to its functionality show potential.
Comments on this post
No comments have been added for this post.
You must be logged in to make a comment.