The latest version of Google’s text-to-image model is now available to the masses.
Teased back in May at Google I/O, Imagen 3 is the tech giant’s most advanced AI generator that rivals other models like Midjourney, DALL-E 3, and X’s very uncensored Grok-2. Speaking of which, Google has taken a different approach than Grok-2, which has swiftly become known for generating copyrighted images and deepfakes of public figures. Unlike Elon Musk’s AI model, Google says it “used extensive filtering and data labeling to minimize harmful content in datasets and reduced the likelihood of harmful outputs.” Images created by Imagen 3 also have Google’s SynthID digital watermark for identifying the image’s provenance.