How to protect yourself in the age of deepfakes

By Rosalyn Page

In the age of deepfakes, a phone call or video clip may no longer be what it seems. 

The term “deepfake” covers AI-generated media such as fabricated videos, cloned voice recordings and doctored images. They can compromise your identity, reputation and privacy through scams, manipulation and social engineering.

Avoiding harm from these fabrications comes down to maintaining an adequate level of trust — something that is built through awareness, crosschecking and maintaining control of our digital footprint.

“You don’t need to live in fear, but you do need to live with intention,” says Nik Kale, principal engineer and AI expert at Cisco.

The biggest misconception, according to Kale, is that deepfakes are a purely technical problem. They are also a psychological one.

“These attacks succeed not because the synthetic media is flawless,” he says, “but because people respond emotionally.” 

Manage your digital footprint

In many cases, deepfakes draw on an individual’s own digital footprint. 

Personal photos, videos, livestreams and long-form audio are the raw material for attackers to create a convincing clone of a person’s face or voice that is then used to deceive someone else.

One approach is to minimise posting high resolution images online, limit access to video recordings and regularly review privacy settings on social platforms to control sharing or access to your chosen friend group.

“Treat your online presence the same way you treat personal information,” Kale says. “The more you expose, the more someone else can use.”

Increase security to help safeguard identity

It is important to actively manage your online presence to limit deepfake risks. Having a consistent, verified presence online ensures others know what you look and sound like. This helps safeguard professional identity and personal security, and “makes it much easier for your network to spot false content,” Kale says.

If a person is in a public-facing role where visibility is important, consider watermarking content or limiting the quality of what is posted publicly.

Setting up notifications such as Google Alerts for your name across major platforms can help catch any impersonation attempts, which can then be reported to the platform. Most major providers, including X, YouTube and TikTok, have dedicated identity theft reporting processes.

“You want to know if someone is using your image or likeness somewhere you did not authorise,” says Nakshathra Suresh, an academic and international AI and cyber safety expert.

With close colleagues or family, establish code words or security questions for sensitive requests.

The advice to parents is to strictly limit images and videos of your children shared online.

“Make those accounts private and be highly selective about who can view content featuring your kids,” says Suresh.

It is also vital that parents teach age appropriate digital literacy to their children and “monitor what they share themselves as they get older,” she adds.

Governments are being urged to legislate to protect people against deepfakes. In Australia, Senator David Pocock has introduced a private member’s Bill that would, if passed, bring removal notices, formal warnings and even civil penalties for people wrongfully depicted or exploited via deepfakes.

How to spot deepfake content

Media literacy begins with assuming that anything unexpected or out of the ordinary could be synthetic. It helps to pause and reflect before reacting, especially when the content is emotional or urgent. 

“Deepfakes often rely on creating panic or surprise, so slowing down gives your brain space to question the situation,” Kale says.

A simple but powerful habit for spotting deepfake content is to always verify unexpected audio or video through a second channel. For example, if someone sends a video asking for money or urgent action, confirm it through a direct text or email. If a strange voice call is received, check with the person directly through another platform. 

“This is called ‘multi-path verification’ and it is one of the most reliable ways to prevent impersonation,” he says.

 

"These attacks succeed not because the synthetic media is flawless, but because people respond emotionally."

 

When viewing any content that seems suspicious, pay attention to things like unnatural eye movement, mismatched speech rhythm, lighting that does not track with the subject or expressions that seem slightly robotic. Look for blurs or distortions, unnatural breathing patterns and the wrong number of fingers or toes.

“You don’t need specialised tools to sense when something doesn’t feel real. Trusting that instinct is important,” Kale says.

Technical verification tools are improving, and many platforms are adopting authenticity and provenance standards that attach cryptographic signatures to real media. These tools can detect tampering or signal when an image or video lacks an original source. 

“They’re not perfect,” he notes, “but they’re becoming valuable for confirming whether content has been altered.”

Trust your instincts — and check the metadata

Context is also important. Unusual requests or calls from “the boss” for help with money should be treated as red flags.

“Similarly, if you receive what appears to be compromising content of yourself that you know you never created, identify and recognise this as a deepfake and report it to the platform, as well as the authorities,” says Suresh.

She also recommends checking the metadata on any video files received, because signs of heavy editing or suspicious origins can suggest it has been manipulated.

If you discover a personal deepfake, document everything before reporting it to the platform and the relevant bodies, for example, Australia’s eSafety Commissioner, which has the authority to direct platforms to remove this content.

Above all, Suresh says, people should trust their instincts and be informed about the dangers. Educated targets are harder to victimise. 

“Understanding that someone can fake a video of you completely changes how you evaluate content.” 

Action points

1. Limit your digital footprint: If possible, post fewer high-resolution images and videos online, and review privacy settings regularly. If you are in a public-facing role where visibility is important, consider watermarking your content.
2. Strengthen identity safeguards: Keep a verified online presence, watermark public content and set Google Alerts for your name.
3. Educate and protect your family: Make children’s accounts private, share only to a trusted group and teach digital literacy as they grow.
4. Verify before acting: Always confirm unexpected requests through a second channel and pause before reacting to emotional content.
5. Spot and report deepfakes: Look for visual or audio anomalies, check metadata and report impersonations to platforms and authorities.