Microsoft Announces Upgraded ChatGPT Model GPT-4 That Supports Visuals

UPDATE: GPT-4 has been launched, and it can now comprehend image prompts. Although significant, the update is admittedly less exciting than previously thought, as news outlets had widely reported that video generation was a possibility. You can read more about the new model here.

Remember when humans had to put in real effort to create something? Well, the game is about to change, again.

When ChatGPT broke into the scene, to say it caused an uproar would be an understatement. No longer did one have to devote hours to composing a thesis, film script, or essay. However, its limitations stop at the fields of text, since this is what its Generative Pre-trained Transformer (GPT) language models, GPT-3 and GPT-3.5, are versed in.

But not for long. OpenAI, the Microsoft-backed artificial intelligence research lab behind ChatGPT, has announced that the soon-to-be-here GPT-4 will be multimodal, which means it will transcend text output to comprehend other forms of media, like images and videos.

Andreas Braun, Microsoft Germany’s CTO, revealed the update at the AI in Focus – Digital Kickoff event, sharing that GPT-4 will arrive with “multimodal models” for “completely different possibilities,” as quoted by German news outlet Heise. Braun named videos as one of these potentials.

In particular, GPT-4 will be capable of translating text, video, images, and even music, elaborated Holgen Kenn, Director of Business Strategy at Microsoft Germany.

Notably, GPT-4 is being touted to support “all languages,” and it can even toggle between languages if that’s what you want it to do. Braun outlined at the keynote that you can ask the model a question in German, and it can reply in Italian.

[via ZDNET and Search Engine Journal, cover photo 267728566 © Rokas Tenys | Dreamstime.com]