• Source:JND
Add as a preferred

Google has been among the top tech companies that changed gears when it comes to AI-based tweaks since last year. Continuing this saga, the company has now introduced ImageFX. The image generation tool will use the Imagen 2 text-to-image model of Google Deepmind to render output. Apart from the new tool, the company said it is improving its other AI tools - MusicFX and TextFX. Here are all the updates concerning the three tools.

ImageFX in Google's experimentation Labs lets people create images using simple text-based prompts. The AI service will also feature "expressive chips" to let users experiment with different dimensions of creativity in a photo. In the example, the company shared an instance of how the chips offer varying choices for a 'Photoreal' portrait to transform it into a 'Hand drawn', Abstract' or 'Impressionist' mode. The tool will likely compete with other AI-based image generation services online, such as Midjourney and OpenAI's Dall-E.

Also Read: Google Maps Is Getting Generative AI Boost To Recommend Places Based On Diverse User Preferences; Testing Now Live In US

Besides ImageFX, the text-to-music experiment - MusicFX can now generate up to 70-second tunes and explore more choices using the expressive chips and download or share your creations. The company is also adding faster output generation capabilities and higher-quality music enhancements.

MusicFX can now generate up to 70-second tunes. (Image:Google)

TextLM, the generative AI experiment catering to "lyricists, wordsmiths", and the ones who wish to explore AI is getting "usability updates" to improve the user experience. The company noted that the technologies will align with the tech giant's AI Principles.

"With the increased ability to generate photorealistic images with Imagen 2, we've made significant investments in training data safety and added technical guardrails to limit problematic outputs like violent, offensive, or sexually explicit content as well as applying filters to reduce the risk of generating images of named individuals," Google said. The tech giant will also conduct "adversarial testing" to curb risks associated with "harmful and problematic" content.

Also Read: Google Bard Update: Gemini Pro In Bard Goes Global, Gets Image Generation Capabilities; All About Updated Chat Model

Output generated using MusicFX and ImageFX will be marked with a "SynthID", a digital watermark into the content for better recognition. Plus, the images will also have IPTC metadata for added identification as AI content. According to the company, these tools will be available in the US, New Zealand, Australia and Kenya. Users can experiment with the features on labs.google.