ChatGPT's next major upgrade could let you generate video from text

6 months ago 7
Person holding a telephone  with ChatGPT open
June Wan/ZDNET

ChatGPT's precocious abilities, specified arsenic debugging code, writing an essay oregon cracking a joke, person led to its massive popularity. Despite its abilities, its assistance has been constricted to substance -- but that is going to alteration soon. 

GPT-4 is acceptable to motorboat adjacent week, allowing for the anticipation of text-to-video generation, according to an code from Andreas Braun, CTO of Microsoft Germany, astatine the AI successful Focus -- Digital Kickoff. 

Also: How does ChatGPT work?

"We volition present GPT-4 adjacent week; determination we volition person multimodal models that volition connection wholly antithetic possibilities -- for example, videos," said Braun according to Heise, a German quality outlet astatine event.

ChatGPT runs connected a connection exemplary architecture created by OpenAI called the Generative Pre-trained Transformer, besides known arsenic GPT, specifically GPT-3 oregon GPT-3.5, which bounds the chatbot to substance output. 

Since GPT-4 would beryllium multimodal AI, the usage cases for it tin spell beyond text, dissimilar anterior models. Holger Kenn, Director of Business Strategy astatine Microsoft Germany, said the AI could construe substance into video, music, and images, too.

Also: ChatGPT can't marque music, but Google's caller AI exemplary can 

A text-to-video generator isn't a wholly caller concept, with tech giants specified arsenic Meta and Google already having models. 

Meta has Make-A-Video and Google has Imagen Video, which some usage AI to nutrient video from idiosyncratic input. However, some Meta's and Google's exertion remains successful the probe signifier and unavailable to the public, placing OpenAI successful the cleanable presumption to enactment up successful the generative AI space. 

Editorial standards
Read Entire Article