1. Leading AI chatbots can generate images in addition to stories, poetry, and code.
2. ChatGPT, Gemini, and MetaAI act as middlemen between users and AI image models for image generation.
3. ChatGPT emerged as the winner in an experiment testing the AI’s ability to create various images based on prompts.
Some of the leading artificial intelligence chatbots, such as ChatGPT, Gemini, and MetaAI, are capable of generating images in addition to short stories, poetry, and code. While the chatbots themselves do not create the images, they act as intermediaries between the user and different AI image models. This approach has led to some challenges, such as Gemini generating racially biased images.
In a series of prompts to test the image generation capabilities of these AI chatbots, the author instructed each bot to create different types of images without providing specific instructions beyond the concept and style desired. ChatGPT, MetaAI, and Gemini were tested on prompts such as a surreal landscape, an old wizard, a cyberpunk ninja, a cute baby elephant, nature and technology, a simple still life, and heading to space.
Overall, ChatGPT was found to be the most versatile and consistent in generating images that aligned closely with the prompts. While MetaAI and Gemini also produced compelling images, ChatGPT’s wider feature set and ability to generate a range of image orientations and canvas sizes gave it a slight edge over the other two chatbots. Additionally, MetaAI was noted for its ability to animate images, offering a different creative dimension compared to the other bots. Each chatbot demonstrated unique strengths and styles, showcasing the AI’s evolving capabilities in image generation.