OpenAI, the developer of ChatGPT, unveiled Wednesday DALL-E 3, the latest version of its artificial intelligence (AI) model specializing in generating images from textual descriptions, intensifying a race among multimodal AI solution providers that offer image creations based on generative AI.

One of the most significant features of DALL-E 3 is that users can integrate it with ChatGPT. If users provide textual prompts, the platform immediately transforms them into images.

OpenAI explained that users can create art without having to put in much effort to craft prompts themselves.

DALL-E 3 also incorporates safety measures to ensure that explicit or offensive images and those violating copyrights are not generated.

In collaboration with external experts, OpenAI trained its language model to disregard specific words or the names of public figures.

DALL-E 3 will initially be available as part of paid services such as ChatGPT Plus and ChatGPT Enterprise starting next month.

DALL-E 3 faces competition with existing solutions such as Stable Diffusion, created by Stability AI, and Midjourney Inc.’s image creator tool with the same name as the company.

Midjourney’s specialty lies in generating images through the messenger program Discord. It has already amassed a significant user base with around 15 million subscribers and approximately 100,000 daily users.

Stable Diffusion stands out for being open source, allowing businesses to utilize it for creating their own services and even commercializing them. The latest edition, Stable Diffusion XL, showcases high quality. Amazon.com Inc. also unveiled on Thursday a new version of its AI voice assistant, Alexa. By applying a large language model to existing speakers, Amazon has introduced a significantly improved voice AI.

Amazon plans to incorporate generative AI into new products featuring Alexa.

By Lee Duk-joo and Chang Iou-chung

[ⓒ Pulse by Maeil Business News Korea & mk.co.kr, All rights reserved]