Image from 'ChatGPT is not all you need. A State of the Art Review of large Generative AI models' (Gozalo-Brizuela and Garrido-Merchan, 2023)
The graphic above covers the following AI models. Although it is somewhat out of date (it was written when ChatGPT was still in version 3.0), it provides an idea of the types of AI models aimed at the public and some examples of each.
For details, reference the related paper (Gozalo-Brizuela and Garrido-Merchan, 2023).
While the above graphic and related pre-publication paper are likely to age quickly, both provide a snapshot of some of the different AI generative AI models and the types of output possible.
Besides these, of course, there are the voice-to-text, text-to-voice AIs used for years for auto-captioning of videos and for reading text aloud to the sight-impaired, as well as that used by personal assistants such as Siri and Alexa, plus Google's audio search. Additionally, Google has image recognition search functionality and both audio and voice search functionality.
Depending on the platform, AI can take in text, code, images, animations, video and audio and output the same or different formats. Image generators can take in text and create images based on millions of images from its database. If you want to make the image 3D, you can find software that will help you do that. Want to animate that 3D image? There are programs using AI that can do that. Want to create a voice-over without a person? AI has been doing that for years for sight-impaired people and is now available to map to a 3D avatar to create the illusion of a human video. Video producers are experimenting with AI to create entire short films with script writing, characters, backgrounds, music and narration produced by AI with minimal human intervention. See some examples in this YouTube playlist.
While in late 2022 to early 2023, people were making fun of the results, technologies are learning and improving rapidly.