What Is a Deepfake? How to Recognise One and Protect Your Business

Picture a voice message that sounds exactly like your best client, asking you to change their bank details before the end of the day. Except it isn't them. That's the moment deepfakes stop being a tech curiosity and become a business risk.

This guide explains what a deepfake is, how they're made in broad terms, how scammers use them, and the practical checks you and your team can use to stay safe. I won't explain how to create one, because that isn't something anyone running a business needs.

What a deepfake is

A deepfake is an image, video or piece of audio that's been edited or generated with artificial intelligence so that it looks or sounds like something real. The word blends 'deep learning', a type of AI, with 'fake'. According to Wikipedia, the term came from a Reddit user called 'deepfakes' in late 2017, who shared face-swapped videos.

Today the word covers a lot of ground: a video of a celebrity saying things they never said, a cloned voice on a phone call, a fake photo of an event that didn't happen, or a real recording with a few words quietly changed.

How deepfakes are made, in broad terms

You don't need the technical detail, but it helps to know why deepfakes have got so convincing. They're built with neural networks, AI systems that learn patterns from large amounts of examples. One common approach uses a network called an autoencoder that learns what a face looks like, and a later upgrade adds a second network whose job is to spot fakes, so the first one keeps getting better at fooling it.

The practical upshot is simple. The more photos, videos or voice clips of a person are available online, the easier it becomes to imitate them. Business owners who post a lot of video, podcasts and talks have more raw material out there than most.

How scammers use deepfakes against people and businesses

Deepfakes aren't only about viral videos. These are the uses that matter most for everyday people and small firms:

• Voice-clone calls. The FTC warns that a scammer can clone a loved one's voice from a short audio clip and call with a fake emergency.
• Fake boss or client requests. A cloned voice or video asks a member of staff to move funds, change bank details or share passwords.
• Fake endorsements. Wikipedia notes fraudsters use deepfakes of celebrities and politicians to push fake investment schemes.
• Fake reviews and testimonials. AI-made faces and videos can be used to look like happy customers.
• Misinformation and harassment. Fake images or videos can damage a person's or a brand's reputation.

The same healthy suspicion helps with other fakes too. My guide on telling a good AI consultant from a fake one uses a similar mindset.

How to spot a deepfake video, photo or voice

Spotting a deepfake by eye or ear is getting harder, and Wikipedia's summary of the research says human judgment alone isn't reliable. Still, these signs are worth noticing:

• Lighting or shadows that don't match the rest of the scene.
• Lip movements that drift out of sync with the words.
• Odd blinking, stiff body language or a face that looks too smooth.
• A voice with a strange rhythm, missing breaths or an unusual flat tone.
• A message that creates urgency and asks you to skip your normal process.

The last one is the most useful. A deepfake might look perfect, but the request behind it usually feels wrong. Urgency, secrecy and a new way of sending funds are the real warning signs.

Deepfake checks to build into your business

The best defence isn't spotting the fake, it's having a process that a fake can't get around. Here's what I'd put in place:

1. Agree that any request to move funds or change bank details is confirmed by calling back on a number you already have, never the one in the message.
2. Set up a simple code word with family members and key colleagues for real emergencies.
3. Require two people to approve unusual or large transfers.
4. Tell your team that it's always fine to pause and check, even if the request seems to come from you.
5. Keep a short list of how each client, supplier and bank normally contacts you.

The FTC's advice boils down to one line I love: don't trust the voice. Hang up and call the person back on a number you know is theirs. It also points out that requests to send cryptocurrency, use gift cards or wire funds are classic scam signs.

What to do if you're targeted by a deepfake

• Stop and don't reply to the message or call.
• Contact the real person through a number or address you already trust.
• If you've sent funds or details, contact your bank straight away.
• In the US, report scams to the FTC at ReportFraud.ftc.gov.
• In England, Wales and Northern Ireland, the NCSC says to report fraud to Report Fraud; in Scotland, report it to Police Scotland.
• Report fake videos or images of you to the platform hosting them. If you're in danger, contact the police.

Deepfakes, AI images and honest marketing

Not every AI-generated image is a deepfake. Making an illustration for a blog post or a product mock-up is a normal use of AI image tools, and my guide to AI images for business covers that side. The line is crossed when content is made to look like a real person or a real event without consent, or to deceive.

For your own marketing, a few habits keep you on the right side of that line: only use photos of people who've agreed, label AI-generated visuals where a platform asks you to, and never use AI to imitate a real person's face or voice without written permission.

Why deepfake awareness belongs in staff training

Most deepfake scams succeed because someone feels rushed and alone. A ten-minute chat with your team about voice clones and fake requests can stop a painful mistake. If you'd like help writing simple AI and security guidelines for your business, it's part of the work I do as an AI consultant for small businesses.

Frequently asked questions

What is a deepfake in simple terms?

A deepfake is an image, video or audio clip that's been edited or created with AI so it looks or sounds like a real person or event when it isn't.

Are deepfakes illegal?

It depends on what they're used for and where you live. Using them for fraud, harassment or non-consensual imagery can break the law. This isn't legal advice, so check the rules for your situation.

How can I tell if a voice call is a deepfake?

You often can't by ear alone. Hang up and call the person back on a number you already have, and be wary of urgent requests to move funds or share details.

Where do I report a deepfake scam?

In the US, report it to the FTC at ReportFraud.ftc.gov. In England, Wales and Northern Ireland, use Report Fraud, and in Scotland contact Police Scotland. Also tell your bank if funds were sent.

Can small businesses be targeted by deepfakes?

Yes. Scammers can impersonate an owner, client or supplier to request transfers or bank detail changes, so a call-back rule protects you.