Google is rolling out a series of updates on the generative ai front, including a new text-to-image conversion tool. What's different ImageFX is that it has an interface that features “expressive chips.” The idea here is that these will help you “quickly experiment with adjacent dimensions of your creation and ideas.”
In addition to the debut of ImageFX, Google says it has improved MusicFX and TextFX. The company says it has made updates to the MusicLM model that include faster music generation and higher quality audio, along with new features. Generated songs can now be up to 70 seconds long. as for TextFXGoogle has implemented usability updates with the goal of improving navigation and the overall user experience.
Images generated by ImageFX and audio created with MusicFX are tagged with SynthID, a digital watermark that aims to make it clear that they are faked using artificial intelligence, especially when they appear in Search or Chrome. ImageFX creations will also include IPTC metadata. This, according to Google, will offer “people more information every time they find our ai-generated images.”
People in the US, Kenya, New Zealand, and Australia can try these new and revamped tools in the ai Test Kitchen starting today. For now they are only available in English.
The Image 2 model powers ImageFX's new imaging capabilities. It's also the technology that's powering new generative ai options in Bard, Search, Ads, Duet ai in Workspace, and Vertex ai. Google says Image 2 helps deliver the highest quality ai-generated images yet. The company notes that the model helps keep images free of artifacts and improves areas of imaging that these tools have struggled with until now.
Additionally, Google says it has made “significant investments” in the security of Image 2 training data while adding guardrails to “limit problematic results like violent, offensive, or sexually explicit content, as well as applying filters to reduce risk of generating images of named people”. “This is due to the model's improved ability to generate photorealistic images. The company says it also conducts “extensive adversarial testing” to detect and clamp down on potentially problematic and harmful content.
Elsewhere, the Gemini Pro in Bard is more widely available starting today. Is now accessible in more than 40 languages and spanning 230 countries and territories. Plus, starting today, Google says people in most countries can generate images in Bard in English for free. These images will include SynthID watermarks.