> ## Content Index
> Fetch the complete content index at: https://nordisk-ai.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# Microsoft VASA-1: An AI Model With Potential for Deepfake Abuse
- URL: https://nordisk-ai.ghost.io/microsoft-vasa-1-an-ai-model-with-potential-for-deepfake-abuse/
- Published: 2024-05-24T11:44:44.000Z
- Updated: 2026-08-27T17:21:47.000Z
- Description: Microsoft's new AI model VASA-1 can create videos from just a portrait photo and an audio recording, sparking both fascination and concern over potential misuse in deepfake technology.
- Author: Anders Hagen
- Tags: AI Models, AI Model, Cybersecurity, Deepfake, Microsoft, security, StylaGAN2, technology, VASA-1, #Migrated-1787841634512, #wp, #wp-post, #Import 2026-08-27 14:41, #Import 2026-09-07 13:41

Microsoft has recently launched a groundbreaking AI model, VASA-1, which can create a video using just a portrait image and an audio recording as input. The name choice, however, might come across as unfortunate in the Swedish market.

![](https://storage.ghost.io/c/ae/3a/ae3ae746-0116-4fb1-8b39-8b60b897bca8/content/images/2026/08/microsoft-deepfake-ai-modell.jpg)

## Advanced Technology for Video Production

VASA-1 doesn't require an actual photo to work. During the demonstration of the AI tool, VASA-1 was shown making the Mona Lisa sing. Not all the portraits used in the demonstration were of real people; some were created using the AI tools StyleGAN2 and DALL-E 3.

## Risks of Deepfake Technology

Despite the impressive technical advances, VASA-1 also raises concerns about potential misuse. With the ability to create realistic videos from simple portrait images and audio recordings, the risk increases that the technology will be used for deepfakes.

[See More About the Model Here](https://www.microsoft.com/en-us/research/project/vasa-1/?ref=nordisk-ai.ghost.io)