Meta's New AI Image Model: Game Changer or Gimmick for Your Ads?

Meta has just launched a new AI image generation model, and I've been putting it through its paces. Honestly, this isn't the same old clunky AI we've seen from them in the past. It looks really good, and it genuinely has the potential to make your ads work better. Previously, Meta offered an AI creation module within the ad launch process, but let's be frank, it was horrible. This new iteration, however, feels entirely different and far more sophisticated.

Currently, the model is live on Instagram, and you can already start experimenting with it for your organic content. From my testing, one of the most significant improvements is that it has finally solved the persistent issue of product distortion. This was a major hurdle with previous AI tools, often rendering generated images unusable for ecommerce. Now, products appear crisp and accurate, which is a massive win for advertisers.

The vision for the near future is clear: you'll simply upload a catalog of your products, and Meta's AI will generate all the necessary creatives for your campaigns. This promises to be a huge time-saver and a powerful tool for scaling ad production.

The Catch: AI Labeling

However, there's a critical catch we need to address. Every single creative that Meta's new AI model produces is automatically labeled as "AI" in the top right corner. This isn't a minor detail. When potential customers see that an image is AI-generated, their propensity to purchase drops significantly. In theory, this would mean you shouldn't run any AI-generated ads if you want to maintain conversion rates.

Navigating the AI Landscape

The reality, thankfully, offers a workaround. If you create an image using an external AI tool, for example, Midjourney, DALL-E, or any other third-party generator, and then upload it to Meta, the platform cannot reliably detect that it is AI-generated. They simply don't label it. Our data shows that Meta only manages to correctly identify and label approximately 14% of externally generated AI ads. This presents a clear strategic advantage for advertisers.

What this means is that while Meta's new internal tool is going to be incredibly useful for rapid prototyping and generating ideas, it won't necessarily be the primary source for your final ad creatives if you want to avoid the "AI" label. It's a powerful addition to your toolkit, but it isn't going to replace your creative team entirely. Instead, think of it as a facilitator, allowing your team to focus on strategic execution and refinement. We can use it to quickly test concepts, understand what resonates, and then produce high-converting, unlabeled creatives using external AI tools or human designers.