In a brand new podcast, you can hear Joe Rogan interview Steve Jobs, the co-founder of Apple. They chat for nearly twenty minutes, which is impressive considering Jobs has been dead for over a decade. But this is not a long-lost recording: Joe Rogan has never interviewed Steve Jobs - and he's not doing so here, either. The interview is entirely false, generated as a demonstration of what technology can do. It's a deep fake.

Deep fakes are becoming more common. So far, most examples are novelties - you can see Sylvester Stallone as The Terminator or watch Barack Obama say things he'd never utter. Admittedly, both show signs that they are not real, and even the Rogan/Jobs interview has a fakeness to it. But experts predict that near-perfect deep fakes are on their way.

This poses a big challenge for the rest of us. Seeing used to be believing, but now we can't just trust our eyes or even ears. So, how should we spot a deep fake?

How are deep fakes made?

Deep fakes first came into being for more legitimate reasons. You might remember the brief fad of recreating dead celebrities such as Tupac Shakur and Michael Jackson. When ABBA recently announced a new album, they also said they were planning live concerts featuring holograms of their younger selves. But not all examples are for entertainment. The Dimensions in Testimony project uses deep fake technology and holograms to capture the stories of survivors from the Nazi death camps.

Still, this same technology could create very deceiving material, such as revenge porn, where the victim's face is superimposed over a pornographic video or photo. Celebrities, in particular, are often targets for this type of fakery. There is also the real possibility that deep fakes can become propaganda. Indeed, it's already happening, such as a fake video of Ukraine's President saying they will surrender to Russian invaders.

Fortunately, the quality of the fake was not very good. But they are getting better and don't need much to do it. Artificial intelligence can study photos and videos of a person, capturing their likeness, expressions and voice. Then it's only a matter of giving the AI a new script, and it can generate a fake video, photo or audio clip. You can even try it yourself - there are many different deep fake services and apps. Most produce work that is obviously fake. But the really advanced stuff is difficult to spot.

How to spot a deep fake

It begs the question: how can you spot a deep fake? The answer is, unfortunately, that it's not always easy. With poorer-quality videos and photos, you can usually see or hear that something is off. Common signs are mismatched lighting, lip-syncing that sometimes isn't accurate, and blurring around the person's neck and jawline. Another trick is to slow the video down - since video deep fakes are created frame-by-frame, you can sometimes spot glitches. There are more ways to look for such manipulation - MIT lists eight ways to spot a deep fake.

But if the fake is really good, especially on an image, it's nearly impossible to spot unless you know what to look for and study the image closely. In this regard, digital fakery has been around for a long time - magazines and ad agencies have been altering images with Photoshop for decades, often to the degree that you cannot see the changes. While few have stooped so low as to put someone's head on a different body, it does happen - we often don't realise it.

Experts are working on the issue. Projects such as Reality Defender hope to spot deep fakes for you one day. But for now, it's on us to make the distinction through some deduction: do you trust the source, and does the content seem plausible? It's essential to be objective - you might really like the idea that Joe Rogan interviewed Steve Jobs. But you know that's impossible. So beware of confirmation bias - the tendency to accept information just because it confirms your opinions or desires.

But the ultimate test is whether there is a consensus on the material. Someone promoting a false narrative might insist a video is real, but if dozens of others are noting it's fake and showing why it's fake, you're better off following the majority opinion.

Ultimately, deep fakes might force us to be more critical about what we see and hear online. In an age where fake news and social media propaganda constantly try to influence our views, it's a good time to develop our nose for the truth. If it's too good to be true, it's probably fake.

Stay safe

While we may not be totally able to escape online fraud and various digital scams, not letting one's guard down and knowing how to act in the aftermath of being a victim is vital.

Click here for more tips on how to avoid falling victim to digital fraud.