Richer, more engaging lifestyle images

Since we first announced asset generation in Performance Max, we’ve continued to grow and improve our technology so that you can generate high-quality assets that resonate with your customers and deliver strong results. Last year, we expanded asset generation to six new languages, introduced it to new campaign types, including Demand Gen, and upgraded our image generation model.

With the help of Imagen 3, we’re now rolling out the ability to use text prompts to generate images that contain adult people and faces across Performance Max, Demand Gen, Display and Apps campaigns. We’ve conducted extensive user research and developed protocols and safeguards to meet advertisers’ high expectation for quality, while adhering to Google’s rigorous product and design principles and clear Ads policies. For example, you won’t be able to generate images of brand-named products, prominent figures like politicians and celebrities, children or minors, or other sensitive or explicit content. All generated images are tagged with SynthID to provide transparency into which images were generated using AI.

This new capability means that you can create compelling lifestyle imagery to better engage customers. For example, if you’re an instructor who wants to drive leads for cooking class enrollments, you could use a simple prompt like “person cooking” or be more specific with “middle-aged man chopping carrots.” If the initial images you get don’t meet your expectations, you can generate more variations or add more descriptive details to the prompt like age, gender, race, ethnicity and nationality to customize further. If you don’t like the images generated, you can “Hide” them and provide feedback to help our AI improve. From start to finish, you’re in control of the creative process and able to approve every image before it goes live.