Video Gamer is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Prices subject to change. Learn more
If you were wondering what Visual ChatGPT is, don’t worry! We’ve got you covered.
ChatGPT is an AI chatbot that has been taking the world by storm since its release in November 2022. It’s not surprising given that ChatGPT can write articles, song lyrics, poems, and even code with only a few prompts to get it started. Even separate from how useful it could be in the future, frankly it’s fascinating to see an AI manage all those tasks so convincingly.
Now the latest news is that GPT-4 will be arriving this week and Visual ChatGPT will follow soon after.
But what is Visual ChatGPT? Let’s take a look.
Jasper AI
Copy AI
Originality AI detector
What is Visual ChatGPT?
Visual ChatGPT is Microsoft’s new AI. It combines existing chatbot technology with VFMs to allow users to chat via text and image. Visual ChatGPT will also allow users to create and edit images based on user prompts. This makes Visual ChatGPT a multimodal AI and exciting step forward for ChatGPT.
Let’s have a look at how they’ve managed it.
The Visual ChatGPT uses the existing chatbot framework of ChatGPT with VFMs.
A VFM is a Visual Foundation Model. (A Foundation Model is essentially an AI framework through which it learns from massive datasets and produces specified content. You can find more information about Foundation Models on the NVIDIA website here.) A Visual Foundation Model is exactly the same as a Foundation Model except designed for use on images. VFMs are all carefully programmed to do very specific tasks to the best of their ability.
By combining multiple VFMs like Stable Diffusion and Transformers, and building them into the existing ChatGPT framework, the Visual ChatGPT aims to be greater than the sum of its parts, allowing for a variety of editing capabilities and image creation. For example, using Visual ChatGPT you could create an image of a flower based on an existing image, however with new colors and in a cartoon style!
Read More: ChatGPT alternatives in 2023- free AI writing tools
How is Visual ChatGPT different to ChatGPT?
Aside from the obvious introduction of images generation and editing, the main difference between ChatGPT and Visual ChatGPT is the introduction of the Prompt Manager into the AI framework.
The Prompt Manager carries out a series of functions:
The first is to inform the ChatGPT framework what the VFMs actually individually do.
Then it converts the data provided by the various VFMs into a format the chatbot technology can understand.
Finally, the Prompt Manager runs damage control. With the sheer number of VFMs included in the Prompt Manager (twenty-two to be exact), there’s bound to be conflict between command order, command information and all the fiddly stuff that goes into making an AI work. The Prompt Manager smooths all that over so the whole Visual ChatGPT can function as a cohesive unit. If you want to check out the nitty gritty method behind the madness, head on over to this paper submitted by Microsoft outlining how the Visual ChatGPT was made and tested.
Interestingly, in that same paper the creators point out a flaw they have already detected in their AI. Microsoft want to add a self-corrective module to the Visual ChatGPT. This is a smart idea as it will allow the AI to improve itself and learn more as it is used.
Read More: ChatGPT cost – do you have to pay?
Hopefully, we will see this in the next iteration of Visual ChatGPT. For now, we’re just excited to see what Microsoft have got for us already and how Visual ChatGPT will transform the AI world.
Frequently Asked Questions
Who owns ChatGPT?
OpenAI is a San Francisco based company that owns and developed the AI ChatGPT. Microsoft have a multi-billion investment in OpenAI.
What does ChatGPT stand for?
ChatGPT is an acronym for Chat Generative Pre-trained Transformer. ChatGPT is the name of an AI chatbot that generates pieces of writing based on user given prompts.
