I went to a talk the other day on social media and the guy giving the talk, who was promoting his book, made quite a few negative comments about it, describing some of the dangers and the way in which it has evolved into something which its creators didn’t initially intend. In some ways I could see what he was saying although I try to take a more positive view generally. However after the talk I wrote up an essay which I posted on this website and then shared it through social media. I had really very little response and it made me think “what use is this?” Another thing that the author didn’t mention, although he may have covered it in his book, is the way in which computer learning, that is, so-called artificial intelligence, has developed over the last quarter of a century. He did talk about robots and they are closely linked. However the rise in computer learning is possibly not recognised by a lot of people and its relationship to social media.
When I was at Uni in the 90’s I studied linguistics and I did a module called Introduction to Cognitive Science. At the time the field was relatively new and developing rapidly. I was interested in it because it was a multidisciplinary field which combined many subjects which I was interested in including Linguistics, Neuroscience, Computing, Psychology and a few others. The hot topic was neural nets and how computers could be set up to mimic the way that the brain works by using networking and then utilising the enhanced computing power of the network start to learn in the same way that humans do. Another hot topic was speech recognition and it was still in its early stages. Come forward to our present time and it can be seen that the advances in speech recognition have been remarkable. Siri, Alexa and all the other speech based computer interactions clearly show that. What is not so obvious is how this is achieved. The system “learns” how to recognise your speech patterns and then “learns” what sorts of responses to give. This is not programmed in as such but is independent of human interference and is based on the “learning” made possible by massively increased computing power and memory storage. You might have even interacted with computers whilst thinking they were people if you have used chat based support online. Their responses seem quite natural and relevant and therefore you could be tricked into thinking they are human.
So how does this relate to Social Media. Well although we know that we are communicating with other people through social media it is possible that machine learning is happening based on our interactions with it and that the platform itself has learned in a way to censor the interactions, mainly by its curating of our timelines. You can of course still go and look at someones profile and see all their posts but when you look at your timeline that is not necessarily a representation of all the posts of the people you are wanting to get notifications from. There are ways of changing the settings on this to see posts chronologically but more often than not people don’t use this capability. So if the platform is “learning” from the interactions we have, which also includes the language, because it has the ability to break down the language into chunks it can analyse, then what are we teaching it? And then what responses is it making to the interactions we have with it? I’m now starting to think about this quite carefully. And I have to say I’m starting to think about shying away from social media.
See a full list of my blog posts here.