Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

Microsoft teases deepfake AI that's too good to release

Apr 20, 2024 - theregister.com
Microsoft has demonstrated VASA-1, an AI framework that can create videos of people talking from a still image, audio sample, and text script. However, the company has decided not to release the technology due to ethical concerns, as it could potentially be misused for impersonating humans. Microsoft's researchers suggest that the technology could have positive applications, such as enriching digital communication, increasing accessibility for those with communicative impairments, transforming education methods with interactive AI tutoring, and providing therapeutic support and social interaction in healthcare.

Despite this, there are concerns about the misuse of such technology, with the US Federal Trade Commission previously warning about AI-generated videos being used for impersonation fraud. Kevin Surace, Chair of Token, a biometric authentication business, expressed doubt about the viability of restrictions on the technology, pointing out that there are already open source models that are similarly sophisticated. Meanwhile, countries around the world, including Canada, China, and the UK, are attempting to regulate AI-fabricated people.

Key takeaways:

  • Microsoft has developed a framework called VASA-1 that can create videos of people talking from a still image, audio sample, and text script, but has decided not to release it to the public due to ethical considerations and potential misuse for impersonation fraud.
  • Despite the potential misuse, some experts believe the technology has positive applications such as personalizing mass communication, increasing accessibility for those with communicative impairments, and providing therapeutic support in healthcare.
  • Kevin Surace, Chair of Token, expressed doubt about the viability of restrictions on such technology, pointing out that there are already open source models that are similarly sophisticated.
  • Countries around the world, including Canada, China, and the UK, are trying to regulate AI-fabricated people and deepfakes, with laws being introduced to protect victims of non-consensual deepfake images.
View Full Article

Comments (0)

Be the first to comment!