A major software brand has introduced a deepfake detector designed for regular users


The fight against deepfakes will be a key challenge in the coming years. — AFP Relaxnews

Faced with the onslaught of deepfakes on the Internet, publisher McAfee has announced the imminent arrival on PCs of its first detection tool specifically designed to recognise these realistic-looking AI-generated videos that can deceive many of us.

A deepfake usually takes the form of an ultra-realistic-appearing fake video, produced with the help of generative artificial intelligence. For example, it’s possible to superimpose one face on another, or have someone say words they’ve never actually uttered. This also applies to audio files, which reproduce the voices of celebrities to make them say anything.

And these creations add to the growing piles of misinformation – and disinformation – proliferating on social media. And for the people whose image or voice has been used in these fakes, it's tantamount to identity theft. Sometimes deepfake technology is used in the production of humorous content, but these cases are also rarely without the agreement of the individuals concerned.

The greatest risk observed in recent months relates to manipulation of public opinion, with the aim of influencing elections or triggering conflict. Deepfakes therefore can present a major risk, especially when they feature political figures.

While at first deepfakes were relatively easy to spot (irregular movements, suspicious shadows, sound shifting in relation to mouth movements, etc), these days they have become so sophisticated that it's almost impossible to tell the difference between an original video and a totally “fake” one. So Internet users will likely need to increasingly turn to various tools to help detect such content.

Anti-virus specialist McAfee now offers a tool called Deepfake Detector to help people spot this type of scam and avoid being swayed by false information. This AI-powered technology can analyse videos in seconds and alert you if it detects computer-generated content. It will be available, exclusively, on selected Lenovo PCs by autumn, initially in the US, UK and Australia. An annual subscription is required (US$9.99 in the US).

Developed in partnership with chipmaker Intel, the software, which runs in the background, actually harnesses the computer's processing power to analyse these videos directly on the device, bypassing the cloud. McAfee claims an accuracy rate of 96%.

This is one of the first major consumer solutions offered by a major brand. Honor unveiled an on-device AI Deepfake Detection system in June at MWC Shanghai. Meanwhile, there are more specialised tools for professionals and government agencies, such as the Sensity.AI platform. – AFP Relaxnews

Follow us on our official WhatsApp channel for breaking news alerts and key updates!
   

Next In Tech News

Study: Smart devices gathering excessive amounts of personal data
Android: How to get back notifications you accidentally swiped away
How to transfer data from an old console to the Playstation 5 Pro
Missing spooky season? Here are more video games to sate your appetite for horror
Instagram criticised for cutting video quality on less popular posts
Why Google is splitting future Android updates into two-stage releases
WhatsApp rival Signal now lets tidy your chats into folders
Review: With PlayStation 5 Pro, Sony creates attractive alternative to high-end gaming PC
OpenAI's ChatGPT back up after brief outage
Booking.com could cut jobs as part of reorganization plan

Others Also Read