The American tech giant continues to amaze in terms of artificial intelligence. Its new AI tool can thus bring a simple photograph to life.
When we’ve been talking about artificial intelligence for two years, it’s hard not to immediately think of Microsoft. The firm founded by Bill Gates is in fact the privileged partner of OpenAI, the one which allowed it to bring ChatGPT into the world. But the American group is not only present in support, since it is also developing its own AI tools, as we can see with Vasa-1.
Vasa-1, the AI that takes photos
Vasa-1. This is a name that should perhaps quickly become famous if Microsoft’s promises are kept. Because the AI that he has just presented to us displays results that are as fascinating as they are worrying. This allows, using a single photo posted online, to make the image in question move and make it repeat a text, and even a song, to perfection.
As you can see in the example below, the result is quite impressive. The movement of the lips is thus coordinated with the words expressed, while the movements of the head or eyes, as well as the facial expressions, truly give the impression of being faced with a real recording.
A future danger for deepfakes?
Technology that advances in this way is beautiful… but it can also be particularly dangerous! Since the emergence of ChatGPT, a worry has accompanied the development of the technology, which is that ultimately we may no longer be able to distinguish fact from fiction. And this is exactly what Vasa-1 allows with its performance.
A problem of which the team which brought this instrument into the world is aware. Reason why she did not wish to provide “ an online demo, API, product, additional implementation details, or other related offering. » And this is so that malicious minds do not quickly set up scams or disinformation campaigns.
Download
- DALL-E 3 integration for more creative and realistic image creation
- GPT-4 Vision image processing capability for precise contextual responses
- User-friendly interface integrated into various Microsoft products
Microsoft Copilot is a chatbot combining advanced artificial intelligence with the ability to generate creative and realistic images using DALL-E 3, and process image-based queries using GPT-4. This multimodal integration makes it a versatile tool for users looking to obtain contextual information about images or generate tailor-made visual content.
Microsoft Copilot is a chatbot combining advanced artificial intelligence with the ability to generate creative and realistic images using DALL-E 3, and process image-based queries using GPT-4. This multimodal integration makes it a versatile tool for users looking to obtain contextual information about images or generate tailor-made visual content.
Source : Microsoft, Engadget
10