Working from home could be over thanks to deepfake technology and AI that lets you fake video calls

Deepfake might’ve seemed funny at first, when the joke was on Tom Cruise.

But now the explosion of synthetic content and ‘fauxtography’ is rapidly changing how we interact, in our increasingly online lives.

Seeing is no longer believing.

The US election race is rife with digitally manipulated images and the Queensland election also grappled with the same issue.

There are countless deepfakes of Harris and Trump doing the rounds on Reddit and Twitter — all stating they are 100 percent genuine.

Experts say video conferencing could be the new frontier for criminals, to scam businesses who are reliant on technology to do their daily trade.

The prospect of deepfake Zoom calls, where a person believes they are trading commercially sensitive information with their boss but are in fact talking to an online clone controlled by a criminal — could tank work from home for good.

A report by the World Economic Forum found 66 per cent of cybersecurity professionals experienced deepfake attacks in 2022. While 38 percent of large companies have been attacked by deepfake fraud.

Some of the synthetic content is extremely sophisticated, and it’s getting harder to decipher fact from fiction with the sheer volume of content coming at us. If you use social media, you are exposed to more than 3.2 billion images and over 700 thousand hours of video – a day.

There are ways to spot the fakes.

Synthetic videos have their own oddities, like slight mismatches between sound and motion and distorted mouths.

They often lack facial expressions or subtle body movements that real people make.

The absence of blinking could be a sign that the video is computer generated.

However, updated versions of software like Midjourney are rapidly ironing these kinks out.

For those who don’t feel like being duped, there is software — for deepfake detection and deepfake watermarking — but it comes with its own limitations and can’t always be relied on.

Image verification tools like the REVEAL Image Verification Assistant are also handy for detecting mis/disinformation, but you need to be savvy to interpret the results.

Plus, it begs the question - should we be using AI tools to fight AI fakes?

The best defence you have is yourself. Think first and ask simple questions to determine if something is fake.

Quite often your own conclusions will be enough to determine what’s real and what’s not.

Shaun says…

If you are going to experiment with generative AI, here are a few things to keep in mind.

• Privacy: Don’t share personal info with AI software. We know those systems farm our facts for training. Whatever you input, you are handing over to the companies that have created the software for them to use as they choose.
• Purpose: Why are you using the AI software? Are you using it to copy something that has been created by someone else? Are you reusing someone else original work, whether it’s print, artwork or design?
• Consent: If you’re creating an image, who are you depicting? Is it parody? Could they be harmed by the portrayal? This technology can quickly cross ethical or legal lines, especially with unauthorised use of people’s likenesses. Make sure you understand local laws regarding privacy and impersonation and get permission if using real people’s images or voices.
• Disclosure: Many brands are now implementing verification logos but there is a responsibility on individuals as well. If you’re sharing your AI creations on social media, have you made it clear they are computer-generated? What would happen if they were shared further without that disclosure?
• Fact-check: Where is an image from? What is its purpose? Have you cross-checked with other online images to see if it has been deceptively altered. Don’t share images on social media that you think could be altered, especially if the intent of alteration is to deceive.