Artificial Intelligence (AI) is a buzzword that has transcended industry boundaries, becoming a foundational element for major corporations, with one notable exception being Apple. Companies like Microsoft, Google, Meta (formerly Facebook), and even Amazon have been diligently developing their own language models for some time now. In a recent public unveiling, Meta has taken a significant step by introducing “Imagine with Meta AI,” a website that enables users to harness the power of AI to generate images based on textual descriptions. As a rule of thumb, the more detailed the description, the more impressive the results.
Meta employs its proprietary language model called “Emu” for this endeavor, which operates in a manner similar to well-known and widely used models like DALL-E, Adobe‘s Firefly, and Midjourney.
This new AI initiative complements the existing “Reimagine” feature available on both Facebook Messenger and Instagram. Reimagine allows users to respond to conversations with real-time AI-generated images. Imagine with Meta AI, on the other hand, provides users with the same AI image generation capabilities as Reimagine but as a standalone platform. Both applications rely on the same language model, Emu.
How to Create Images with Meta AI
It’s worth noting that although Meta AI initially targets users in the United States, individuals outside of the region can access the service through a VPN, connecting to a server located in the United States. Here’s a step-by-step guide to using Meta AI:
- Access the Meta AI website using the following link: Imagine with Meta AI website
- On the left side of the website, you’ll find a text box where you can input a description of the image you want to generate. On the right side, you’ll see some of the images created by this AI. These images may not be directly comparable to those generated by Midjourney, Firefly, or DALL-E. To proceed, you’ll need to log in with your user account, whether it’s through Facebook or Instagram.
- Each prompt you provide generates four different images. These images come with a watermark in the bottom left corner, which cannot be removed through the website itself (although third-party tools might offer a solution). Click on each image to view it in a larger size.
- To download your created images, hover your mouse over the enlarged image and click on the three horizontal dots located in the upper right corner. Select the “Download image” option, and the image will be saved with the prompt text used as the filename.
In terms of resolution, Meta’s solution offers a reasonably acceptable 1280×1280 resolution. This provides greater flexibility when working with the image, allowing for cropping without significant loss of quality or pixelation in the final result.