Skip to content

There is no Algorithm for Truth

Para-Social

  • People watch videos/tv programs/... not for the content, but for the presenter
  • Example: Twitch streamers
  • Example: Presenters, like David Attenborough
  • Even platforms advertise for this: Patreon example from talk
  • In the TV days, there were gatekeepers (like an editor), but with YT and similar, everybody can upload something. Because people don't check, and just trust authority, misinformation can run rampant

Why do you think that people watch streamers for hours at a time playing a game, they could play themselves? - Certainly not because of the game. Rather is is the para-social relationship between the streamer and the viewer.

Parasocial is defined as a one-sided relationship and they were always a thing. In the TV day, people watched programs, like Top Gear or Planet Earth, because of their authoritative presenters.

But this gets problematic when they ask for money. On platforms, like patreons, creators have the ability to monetise viewers. In return the maybe provide behind the scene videos or pictures. Or other exclusive content. It gives the viewer the impression that they know the creator. What they are doing is, essentially, selling friendship.

These two sentences are directly from a Patreon on how to gain more paying subscribers. The goal is clearly to make creators more human. That they are more like your school friend with which your share intimate moments. But they are not. Viewers think that they know their idol very well, even though, they never met them. And, conversely, the creator probably never met the viewer. There is such a power imbalance that there cannot be a genuine friendship.

If you want more proof that para social relationships are working, look no further than ads on social media. The reason why company spend big bugs on product-integrations and shout outs is because they work. People trust their beloved creators, even if the creator declares the sponsor.

But this, like parasocial relationships themselves, is nothing new. Actors in the 80s had phone lines where fans could pay to send their idol a message. That is quite expesnive for hearing a recorded message on the phone.

Another issue is, that presenters may not be experts in the topic which they present. At least in case of TV documentaries, radio talk shows or newspaper columnist, there was an editor which proof-read stories. So there was some amount of quality control. It was worth to TV stations, like the BBC, to "reuse" loved presenters as it engaged viewers more.

On social media, everybody has a metaphorical microphone, in which they can spout none-sens all day long, consciously or unconsciously. As we have already discussed, corrections are rarely seen by the same number of people who saw the original post. This means, misinformation spreads a lot easier these days.

I do want to mention, that there are of course individuals who source all their content diligently. Channel like Tom Scott, Vertasium and many more do awesome work in the science communication field.

The AI Revolution

One way users and also

Conclusion

Let's quickly recap the presention

First of all, I think, it is really important for everybody to realise that every algorithm has biases. Those can be extremely subtle, but can have enormous ramifications.

Secondly, every recommendation algorithm lays on the spectrum of speech. Ranging from, everything which is legal is allowed, the Nazi Bar, to only opinions a group agrees up on is allowed, the Echo Chamber.

Lastly, Para-social relationships are a powerful and sometimes dangerous tool. It can be leveraged to educate people on topics they otherwise might not have learned. But it is also used to spread misinformation and sell viewers products and "friendships".

Currently, there is no algorithm which tells the truth, even though we sometimes desperately need it. But this means that our brains have to be this algorithm. As such, we have to be aware on how recommendation algorithms work and how parasocial relationships can trick our brain in to believing things it should.

Discussion Questions

  • What are advantages to bias algorithms? It is handy to have search engines which are biased to your views. Say your looking for cat pictures, a search engine can show you only pictures of cats you like.