Is seeing believing? Or could AI mislead you?

Film & Media: Callum Howard, Limitless 

I’m very passionate about video. To me, it’s the evolution of the campfire. Our ancestors would sit around the mouth of the cave, sharing stories of triumph and defeat, and today we share those stories in video form.

But are those stories true?

I’m always asked whether I feel my business is under threat from AI. With the release of generative video software like OpenAI’s Sora 2, is it time to close the doors on our business? The short answer is no. However, AI’s impact on our ability to believe what we see does concern me.

I still believe that video is the most effective way to build trust with potential customers, but this new toolset has certainly opened the door for bad actors to undermine any confidence we can have in what we see online.

It can be overwhelming when we see what AI can do. Are we doomed? I don’t think so. In fact, I have collated some tips and tricks here to help you spot potentially misleading content online, so you can protect yourself and your business from AI-powered scams or misinformation.

Time sensitive

The easiest and most obvious tells come from the date and the duration. The ability to create good AI video didn’t exist pre-2023, so if the video you’re watching was uploaded before then, you’re probably going to be okay. Keep an eye on the duration too. AI videos typically top out at ten to fifteen seconds, so if the video itself, or a shot within it, doesn’t last for more than that, it’s time to be suspicious.

Question what you see

The biggest hurdle AI faces is teeth. If someone in a video is talking, look at their gnashers. Are they one big band of white? Do the teeth change each time the mouth is opened? Are the teeth constantly blurry while the rest of the image is sharp?

Don’t forget to look at the background. AI has a habit of breaking the laws of physics and logic. People, animals or vehicles can often appear or disappear from behind people or buildings. The same goes for the people you see – does a news reporter change each interview? Does the microphone or clothing stay the same throughout?

Motive and verification

Ultimately, the best antidote to misinformation is critical thinking. If something looks too good (or too bad) to be true, then it probably is.

Can you verify it with a second source? If something is newsworthy, it will be reported by more than one outlet. Is this trying to provoke a reaction from you? If something is designed to be shocking, ask yourself if it could be an opportunity for manipulation. Do you know someone involved? If you see a video that appears to be of someone you know, pick up the phone and call them.

Ultimately, AI is here to stay and will only get better, but our ability to talk to each other and think critically will be our best defence against the erosion of trust in what we see online. I’m not ashamed to say that I’ve been fooled myself, but it’s the willingness to learn from what we see that will keep us safe.