Sign in with Microsoft

Chinese owned TikTok has just shown us that it is nothing more than an unregulated attack vector that can spread fear, uncertainty and deception among its users.

Forgetting for a moment that it is one of the world’s largest data harvesters (well, no don’t forget that) it is the attack vector for sad, sick people, whackos, extremists, activists and nation-states to get into our, and our children’s heads.

That clunky, cute, creative 15-second video clip can also contain extremist hate, hardcore porn, snuff clips, drug use, social conditioning, fake news… Need I go on?

TikTok

And it has over a billion active users and 2 billion downloads from Google Play alone. It is the fourth most downloaded app in the iOS Store.

But there is a little more to this highly addictive drug.

It has the highest average engagement rate of any app, including games like PUGB and about double that of Facebook. The average rate is 52.1%, but on sites with fewer than 1000 visitors, it is as high as 144.9%.

Yes, people spend an awful lot of time on it. In fact, the average daily use of TikTok is over one hour, and that has shot up during COVID to over three hours. TikTok says “Five minutes in TikTok is like one hour of real life.”

What you don’t know is this is a hastily developed Chinese app. It took just 200 days from concept to launch of its Chinese version Douyin that is only for Chinese and now Hong Kong use. It has an ultra-sophisticated AI system considered so secret that the Chinese government won’t let Microsoft and other so-called safe harbour suitors have it.

Why the secrecy?

OK here is some healthy speculation from reputable analysts

Harvard Business School says TikTok is using AI to take over the world. Here AI helps creators to build viral videos by content suggestions, editing, music and filters. What is to stop it subtly influencing creators by promoting specific thought pattern or even party lines? There is a growing base of evidence that it is influencing anti-Trump and BLM videos.

TikTok

AI then uses its omnipotent user database to promote a video to the viewer. How does it know what views like? By the way, it recommends videos to you – you don’t have a choice to search a wider database.

We don’t know exactly what is in the database, but we did analyse its privacy policies, and frankly, it can collect as much information as it wishes and can do anything it wants with it. Our findings on how TikTok monetises our data is here.

Let’s just say that by merely using TikTok, it knows your name, gender, location, smartphone type, IP address, preferences (what you like and dislike) and then extends that to all your friends in your contacts database. Oh, and it also aggregates information from Facebook profiles and other third-parties as well as tracking your online web activity. Why does a cute video clip maker need that?

But it goes further. The TikTok design allows it to interpret users’ feelings. You may comment, upload a ‘react’ video, share a video on a social network etc. From this and things like how fast you swipe through the recommended list (are you feeling good, indifferent, or negative), the speed that you swipe away videos, how hard you press on the screen, time of day use etc., it can refine your profile.

Going back to TikTok’s AI – and this is dangerous for content creators.

AI is capable of accurate face and object recognition. So, if you are a creator (or reactor), your facial data is on its file. That is hundreds of millions of faces linked to their smartphone and its location. Want to arrest someone for sedition? TikTok can locate its users faster than any app – as Hong Kong protesters found.

It analyses objects in the background. Let’s say you create in your lounge it knows the furniture type, window position and the bric-a-brac you may have collected. Not that we would ever suggest this information would be used for nefarious purposes, but it does help build up a socio-economic and demographic profile.