OpenAI launched the latest version of its DALL-E photo generator to a select number of testers. The company then integrated the technology into their popular ChatGPT chatbot.
ChatGPT can now create images. They are incredibly rich in detail.
This week, OpenAI, the San Francisco artificial intelligence startup launched a brand upgraded version of the DALLE-E Image Generator to a select group of users. They also incorporated its technology to create ChatGPT, its popular online chatbot.
It is dubbed DALL-E 3, it can create higher-quality images than the previous models of this technology. It also shows an especially good ability for images that contain numbers, letters and human hands, according to the company.
“It is far better at understanding and representing what the user is asking for,” said Aditya Ramesh who is an OpenAI researcher and OpenAI researcher. He added that this technology was designed to provide a better understanding of the English language.
With the addition of the most recent version DALL-E by integrating the latest version of DALL-E to ChatGPT, OpenAI is solidifying ChatGPT as a platform for dynamic A.I., which can create images, text as well as sounds, software, as well as other media itself. Since ChatGPT became a hit in the last year, it has begun a race among Silicon Valley tech giants to remain on the cutting edge of A.I. by making advances.
Bard. Google’s chatbot, named Bard, launched in March, but only to a small amount of users from the United States and Britain. The idea was initially to create a creativity tool to write poems and emails that generate ideas or blog posts and respond to questions using facts or views.
Ernie. Baidu, the search giant Baidu launched China’s first major competitor in the form of ChatGPT during March. The launch of Ernie, which stands for enhanced Representation via Knowledge Integration, turned out to be a complete failure following a promise of a “live” demonstration of the bot was discovered to be recorded.
OpenAI has been offering ways to connect its chatbot to other online services, such as Expedia, OpenTable and Wikipedia. This marks the first occasion that that the company has integrated the chatbot and the ability to generate images.
The DALL-E application and ChatGPT were, at first, two separate applications. However, with the most recent version, the users can use ChatGPT’s service for creating digital images by simply describing what they would like to see. They can also create images using descriptions created by ChatGPT, which further streamlines the creation of art, graphics, and other forms of media.
In a recent demonstration, Gabriel Goh, an OpenAI researcher, demonstrated how ChatGPT is now able to generate precise texts that are later used to create images. After generating descriptions of an image for a restaurant named Mountain Ramen such as the bot created a variety of images from the descriptions in just a few minutes.
The latest version of DALL-E can create images from multi-paragraph descriptions, and it is able to follow instructions that are laid out in fine detail, Mr. Goh said. Like all image generators as well as other A.I. systems, it is also susceptible to making mistakes as well, he added.
Sandhini Agarwal, an OpenAI researcher who is focused on security and policy, stated that DALL-E 3 tended to generate images that were more stylized, not realistic photos. However, she admitted that the model may be stimulated to create realistic scenes, for example, the kind of grainy photos that security cameras capture.
Most of the time, OpenAI does not plan to block potentially harmful content through DALL-E 3. Mrs. Agarwal said such an approach is “just too broad” because images can be harmless or risky depending on the context they are placed in.