Ik kom nog zoveel parels op Twitter tegen. Als je ze gemist hebt, zijn ze vervlogen en als je ze terug wilt vinden, heb je een probleem. Maar ik zie nog dagelijks mooie of goede inhoud langskomen. Verhalen waarvan ik denk: daar leer ik wat van.
Bijvoorbeeld als Thomas Baekdal er even voor gaat zitten, een man met veel kennis en waardevolle media inzichten. In 18 tweets legt hij uit wat het probleem met algoritmes van user generated platforms is. De grote kennis rondom onderwerpen zit vaak in kleine niches, betoogt hij, en die worden niet als belangrijk genoeg ervaren door de systemen.
This is the challenge the tech companies now face. How do we measure expertise, especially for topics that wouldn’t work from a normal user-generated perspective?
Lees zijn thread!
Erwin Blom
|
|
|
Let’s discuss media literacy and algorithms (thread): 1: The key way to understand social channels is that they are ‘user-generated’. Meaning, they are awesome at anything people do themselves, but terrible at anything else.
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
2: For instance, if you want to learn how to restore a Redline Hot Wheels car? Sure, here is one of the thousands of amazing videos about that: https://t.co/LpUuuY7lr9 |
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
3: So anything that people would be good at doing in their garage (or in places that normal people go) works brilliantly. The problem is when you start to look at other things. For instance, how good would people be at discussing ‘The Pope’ from their garage?
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
4: The answer is obvious. They would be terrible at it because they would just be expressing their opinions, and the only people who spend their personal time talking about that would be nutcases. As a result, a search for ‘pope’ on YouTube = nutcase conspiracy theories. https://t.co/qFgxcl8xHy |
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
5: The problem is not really YouTube, it’s that we are looking for the wrong information on the wrong channel. Or, to put it simply, we are asking a user-generated platform for information about something normal people would be terrible at providing.
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
6: So, what we need to learn (and to teach people) is to validate the topic in relation to the channel. People should be asking: “Is this topic something people sitting in their garage would be able to tell me anything about?”
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
7: There are, of course, exceptions, like when a YouTuber creates a very insightful video about something, but in those cases, we can’t look at it as ‘a platform’ but rather in terms of what that single creator is doing. Are they just sharing opinions? Or are they professionals?
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
8: My point is that this is part of the media literacy that we must teach. Platforms like YouTube will never go away because they serve a critical role. For instance, you can’t find any information about Hot Wheels restorations in a newspaper.
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
9: So we need to learn the difference between user-generated and professionally generated and realize that each has value, but also that this value changes depending on the topic.
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
10: This is also true for publishers. Journalists are kind of in the middle. They are neither users-generated nor professionals, so journalist have to find a way to attach themselves to the most valuable part of each.
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
11: My best example of this is @‘s wonderful YouTube channels, where he is attaching his skill as a journalist to connect you with professionals. And the result is some of the highest quality science videos on YouTube: https://t.co/s5DuNESuEf |
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
12: Of course, the platforms also need to get better at distinguishing between these things. When, for instance, people search for ‘The Pope’ on YouTube, it shouldn’t just rank the results that same way as when people search for ‘Hot Wheels restorations.’
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
13: So, the tech companies need to find a way approach professional and user-generated content differently, depending on the topic. But this is a really complicated thing to do, because this is not just about redirecting professional topics to traditional publishers.
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
14: In many cases, the best professional content comes from the smallest niches, where you have someone truly knowledgeable about a topic and who cares about it passionately. And suddenly were are talking about a very different type of algorithm.
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
15: This algorithm is not based on who has the most views, watch-time, engagement or shares. It’s based on who has the most expertise … but how would we measure that?
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
16: This is the challenge the tech companies now face. How do we measure expertise, especially for topics that wouldn’t work from a normal user-generated perspective?
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
17: In other words, how do we make sure that we don’t end up listening to people’s opinions on topics when we really need a professional voice?
|
|
10:44 AM – 3 Apr 2018
|
|
|
|
|
18: And this question is not unique to YouTube or Facebook. We must ask the same question as newspapers. When you cover a story, who do you interview or bring into the studio? And how do we turn this into a form of media literacy across society as a whole?
|
|
10:44 AM – 3 Apr 2018
|
|
|
FMT
op 04-04-2018
