Google is introducing a number of innovations in the field of generative artificial intelligence, including the new ImageFX text-to-image tool. Along with the tool’s debut, Google said that improved the MusicFX and TextFX tools. The company claims to have performed improvements to the MusicLM model, which include faster music generation and better sound quality along with new features. Generated tracks can now last up to 70 seconds. When it comes to TextFX, Google rolled out usability updates with a goal improve navigation and overall user experience.
Images generated with ImageFX and sounds created with MusicFX are marked with a SynthID digital watermark, which makes it clear at first glance that they are created by artificial intelligence (which is especially handy when they appear in search or in Chrome). Works created with ImageFX will also contain IPTC metadata.
New image generation features in ImageFX drive model Image 2. It’s the technology that powers new generative AI features in Bard, Search, Ads, Duet AI in Workspace, and Vertex AI. Google claims that Imagen 2 provides so far the highest quality images generated by artificial intelligence. In addition, Google performed great security measures, so as not to generate images with violent, offensive or sexually explicit content. It also reduced the risk of generating real persons.
For now, users in the US, Kenya, New Zealand, and Australia can try out these new tools as part of the AI Test Kitchen. So far they are only available in English. Later they will be extended to other countries as well.