cloudThing logo in white
Menu open icon
email: info@cloudthing.com
tel: +44 (0) 121 393 4700
Menu closed icon

General

Generally useful pages

Sectors

We know loads about this stuff

What we do

The Building Blocks for cloudThing Magic

Microsoft Goes To War On Deepfakes

Sun Sep 06 2020

Microsoft have released proprietary tech to help detect deepfake videos

Microsoft have released several new pieces of software to help detect DeepFakes in the hopes that they’ll help improve the cybersecurity and trust of online videos and data on the lead up to big events (such as the US elections coming this November) by preventing the spread of false information.

 

Microsoft have called the first tool to be released ‘Video Authenticator’.

With this, people will be able to analyse both images and videos and determine if it’s been edited or manipulated with AI technology.

The tool will then provide a confidence score, letting end-users know how likely it is the image/video has been altered.

In fact, with videos it can provide a confidence score, in real time, for each frame, as the video is playing to let viewers know what’s real, what’s been faked and where it’s been faked!

It works by 'detecting the blending boundary of the deepfake and subtle fading or greyscale elements that might not be detectable by the human eye.'

It was created (Microsoft have said) by using a public dataset from Face Forensic++ was then extensively tested on the DeepFake Detection Challenge Dataset before release.

The second tool Microsoft have released in their war on DeepFakes will allow content creators to provide certificates of authenticity that their content is real.

These certificates are built directly into Microsoft Azure and stay with the content as metadate no matter where it ends up being used,

open quote mark

These new tools will be initially available to political campaigns and media organisations. This will be a long-term effort, but we hope to have an impact in the lead-up to November

Tim Burt – Vice President, Microsoft

close quote mark

In June and July of 2019, Cyber-Security firm Deeptrace found nearly 14,700 deepfake videos floating around online, up over 84% from December ’18.

Most (if not all) cyber security experts are worried about these figures as deepfake videos can easily be used to manipulate the voting public, sow animosity between rival groups or even frame someone for a crime.

Not Quite Ready To Get Back To Work Just Yet?

BUILD THE FUTURE OF THE NONPROFIT SECTOR… TODAY

TEACHING UNIONS SEE OVER 50,000 NEW MEMBERS AMID JOB SECURITY CONCERNS

DATA PROTECTION & ARTIFICIAL INTELLIGENCE: BEST PRACTICE

Contact Us

Name

*

Company Name

*

Email Address

*

Telephone Number

Is there anything else you'd like us to know?

© cloudThing 2020

Sun Sep 06 2020

email iconinfo@cloudthing.com
© 2020 Copyright cloudThing ltd. All rights reserved. Company registered in England & Wales no. 7510381, VAT no. 152340739