Open AI, the brain behind ChatGPT has announced the launch of the next major update of DALL-E. Dall-E is an image generation tool that allows users to create images based on text-based prompts. The first version was launched in January 2021, and now DALL-E 3 will be available to the general public soon.
DALL-E 3: Release Date
DALL-E 3 officially arrived on 20th August, 2023 but it was exclusively for researchers at that time. In October, a new version of the AI Image Generator got a wider release and is now available on Microsoft’s Bing Chat. Only ChatGPT Plus users have access to this feature. When DALL-E 3 will be launched for users outside Bing Chat/nonsubscribers remains to be seen.
Origin of DALL-E
Dall-E is a generative AI technology that enables users to create new images with text to graphics prompts. DALL-E is a neural network and is able to generate entirely new images in any number of different styles as specified by the user’s prompts.
The name DALL-E is a combination of 2 core themes of tech hinting at the goal of merging art and AI technology. DALL is intended to be evocative of famous Spanish surreal artist Salvador Dali while ‘E’ is related to the functional Disney robot Wall-E. The combination of 2 names reflects the abstract and somewhat surreal illustrative power of the technology, that is automated by a machine.
DALL-E was developed by AI Vendor OpenAI and first came out in January 2021. It uses deep learning models with a GPT-3 large language model as a base to understand natural language user prompts and generate new images.
The first generation of DALL-E generated images from text using a technology known as a Discreet Variational Auto-Encoder that was somewhat based on research conducted by Alphabet’s DeepMind division with Vector Quantized Variational AutoEncoder.
DALL-E 2 improved on methods used for its first generation to create high-end and photorealistic images. It works with the use of a diffusion model that integrates data from the CLIP model to help generate a higher-quality image.
Cost of DALL-E 3
DALL-E 3 is available to users via ChatGPT Plus, a subscription service that costs Rs. 1650 per month in India. The subscription includes unlimited access to both ChatGPT and DALL-E 3.
New Features in DALL-E 3
1. DALL-E 3 now comes with brainstorming assistance. It is built from ChatGPT which means quicker prompt refinement and easy image adjustments. Users can consult with ChatGPT to brainstorm new image ideas.
2. DALL-E 3 has better efficiency and faster speed when generating images from text. The goal is to limit the hassle of generating images, by rendering more detailed, expressive, and specific images that are more suited to the user’s requirements.
3. DALL-E 3 can be used directly through Microsoft’s search engine Bing. It is available through Bing Chat rather than being powered by a stand-alone tool. The feature was introduced in October and can be accessed now. It is available only for ChatGPT Plus users who have to pay a monthly subscription.
4. DALL-E 3 focuses even more on its safety protocols, including preventing the generation of images that feature adult, violent, or hateful imagery. To prevent copyright infringement, DALL-E 3 will decline to produce any images that resemble living public figures or directly imitate the style of living artists.
5. DALL-E 3 uses publicly available data and uses it to create all-new images inspired by previously acquired data. Since all artists don’t wish for their data to be used by DALL-E 3, OpenAI offers 2 options for content owners to opt out by either filling out an online form or blocking the GPTBot (Web crawler designed to gather training info) from accessing their site.
How is DALL-E 3 different from its Predecessor DALL-E 2?
OpenAI claims that the biggest change in DALL-E 3 from DALL-E 2 is its ability to understand text prompts better, specifically longer ones. DALL-E 3 has a better safety protocol system and comes with the ability to consult directly with ChatGPT for brainstorming. Let us talk about these differences in detail.
DALL-E 3 understands context better than DALL-E 2
A better understanding of context is one of the highlights of DALL-E 3. DALL-E 2 had a hard time understanding context completely even when specifically prompted as it ignored certain words. DALL-E 3 works much better and offers users with the image they want.
DALL-E 3 remains hand-in-hand with ChatGPT
DALL-E 3 has been built on ChatGPT so these two can be used hand-in-hand and brainstorm better ideas to refine the prompts. ChatGPT generates a unique, tailored, and detailed prompt when an idea is prompted with DALL-E 3. You can even ask ChatGPT to make further tweaks to the results to get the image you want.
DALL-E 3 Images are Yours
Images created by DALL-E 2 did not belong to the user who created it. With DALL-E 3 images you create are all yours. You don’t need permission to reprint, sell, or merchandise them.
Declines mimicking living artists
Mimicking living artists can pose problems like lawsuits, copyright infringements, and more. DALL-E 3 has been trained to decline generating images that mimic the style of living artists. DALL-E 2 can be prompted to mimic living artists, but DALL-E 3 has been coded to avoid any such commands.
Transparency
AI-generated content is becoming more and more common and thus there is a need for transparency to identify such content. OpenAI is looking for ways to help people differentiate between AI-generated images and human-created Images. An internal tool is being designed to determine whether an image was generated by DALL-E 3 or not.
Competitors of DALL-E 3
1. Google SGE
Google has unveiled a new feature that lets you create images from text prompts using its Search Generative Experience (SGE) tool. It is now available as a part of Google’s Search Labs program and uses the Imagen family of AI models to generate images based on your query. The tool is available to users who are 18 or older.
2. Craiyon
It is very similar to DALL-E and is trained on a smaller database than DALL-E, which makes it less precise than DALL-E. Faces look smeared and indistinct on this one. However, it is free and unlimited and the content range is unrestricted in regard to adult or political subject matter.
3. Wombo Dream
Dream by Wombo is a free AI art generator that is available on their website and as an app for Android and iOS. It Specializes in dreamy, surreal images but can create faces and human figures too. One can augment their text input with a mandatory choice of art style on Wombo. The output comes in portrait orientation and the platform is free to use. Nudity and sexual topics are rejected by the app.
4. Adobe Photoshop Generative Fill
Adobe’s famous Photoshop editing suite has also gained an impressive new AI image generation feature that allows you to extend existing pictures and remove distractions. It is intuitive and easy to use and it doesn’t require any payment beyond an Adobe Creative Cloud subscription.
Limitations of DALL-E 3
- DALL-E 3 might not always generate images that accurately match the text description.
- DALL-E 3 might generate images that are not realistic.
- DALL-3 3 might generate images that are biased or offensive.
- DALL-E 3 is limited by the amount of computing power that is available to it. So, it can generate images of a certain size and resolution.
FAQs
1. Can DALL-E 3 images be used commercially?
Yes, you have full commercial rights to all images generated with DALL-E 3. You can also sell, reprint, or merchandise these images.
2. Is DALL-3 free to use?
No, you need to opt for ChatGPT plus a subscription costing Rs. 1650 to access DALL-E 3.
3. What is the resolution of the image produced by DALL-E 3?
DALL-E 3 generates images with a resolution of 512×512 pixels.
You can follow Smartprix on Twitter, Facebook, Instagram, and Google News. Visit smartprix.com for the most recent news, reviews, and tech guides.
