Tuesday, 5 March 2019

Algorithmic singularity in a cold, uncaring world

James Bridle's excellent video on creepy, pseudo-machine generated children's videos on Youtube makes some excellent (and disturbing) points about how artificial intelligence and machine learning is influencing our society at large. Most successful websites work by fabricating an endless flow of 'content' (which is marketing shorthand for... stuff...).  This content is selected based on who you are, your search history, your level of 'engagement' with certain images, your interests, your social status, your demographic, your ethnicity, your cultural capital. If at this stage you are wondering 'how do they know all this?" ('they' being tellingly vague here), the answer is that 'they' do not. A variety of assumptions, shortcuts and conclusions have been made by an increasingly sophisticated, and increasingly successful artificial intelligence.

Brindle documents how this algorithmically created content is essentially used to exploit and upset young children left alone and unsupervised on the internet for financial gain. Yet while this is creepy enough, even more worrying is that because this works, we, as humans are beginning to emulate the way in which A.I thinks and acts. 

Like many sci-fi shows, Humans explores the singularity, or the point where humans and artificial intelligence are interchangeable and indistinguishable. What could be more terrifying that living in a world where we do not know what is real and what is fake? Yet of course all sci-fi is allegorical; it shows us something about our own world and our own time in a way that seems exciting and fantastical.

The reality is unfortunately far less fantastical. We are increasingly unable to differentiate reality from fiction, and arguably we just don't care anymore. Jean Baudrillard's notion of the hyperreal, where a representation is so compelling that it ends up becoming more 'real' than the original, is now a familiar experience for us all. In her mind exploding book Duty Free Art, German artist Hito Steyerl  considers (among many other things) mobile phone photography. Mobile phones are increasingly installed with A.I that can recognise what we are taking pictures of. It can then use the powers of digital convergence to search the trillions of images on the internet in order to essentially make the image that you just took of your cat look more like a cat. Additionally, your cat photo is teaching millions of other mobile phones around the world to create images that are more cat-like. This utterly bonkers situation is indicative of exactly how much artificial intelligence and machine learning have pervaded and altered every aspect of our lives. The photo of your cat is now more real than your actual cat; a hyperreal simulacrum of infinite digital cats.

We do not understand artificial intelligence. It reaches financial conclusions in ways that we never could, and it is doing so in ever more efficient ways. The results are so horrifying because we just do not understand them. Violent and sexualised images of characters from children's franchises, including live burials, teeth extractions and simulated sex acts simply do not belong together. But when combined they create millions of views, clicked on by curious toddlers, and selected by YouTube's increasingly sophisticated 'watch next' algorithm.


We are increasingly living in a world that is far too complicated to make sense of. In his 2016 documentary HyperNormalisation, Adam Curtis explores in exhaustive detail how audiences, governments and media producers are increasingly creating a hyperreal, utopian and virtual existence in order to escape from the increasingly complex issues that affect us. The continued political instability in the Middle east, and in particular the World Trade Center bombing and the subsequent second Gulf War have pushed us online. Curtis goes as far as to argue that world governments have not only exploited but also deliberately created confusion in order to create a world perpetually on the brink of collapse. However another symptom of living in a world too complex to understand is the creation and cultivation of various online bogeymen. The continued sharing of anti-semitic conspiracy theories, accusing powerful Jewish figures of manipulating world politics and finances is a deeply troubling knee-jerk reaction to a world we no longer understand.   

Alex Garland's 2014 sci-fi film Ex Machina, programmer Caleb falls in love with the artificial intelligence Ava. Ava's entire personality is constructed from data, including the profiles of online pornographic performers. Ava is irresistible to Caleb, as a fragile, flawed and ultimately human character. As an audience, we comfortably fall in to the trappings of the genre, and allow ourselves to follow the familiar narrative of the flawed yet beautiful princess being rescued by her unlikely prince. However, Ex Machina breaks all our expectations. Ava is a machine, programmed for one simple goal: to escape from the house where she is being held. Without giving too much away, the ending is shocking because we cannot justify or empaphise with Ava's actions. Yet Ava's solution to her problem is efficient, highly sophisticated, horrifying, and completely unexpected. Our relationship with online media is like Caleb's relationship with Ava. It is compelling, fascinating and we think that we understand it. Yet it's relationship with us is one that we can never possibly comprehend.

While online media has arguably made our lives more uncertain than ever, there are a few straightforward conclusions that we can take from these thoughts and examples:

  • Online media is hypermodal. It goes beyond traditional mono/multimodal forms of media, allowing the audience at least the illusion of control and interaction
  • Online media is hypersegmented. It targets millions of tiny audience groupings on a massive scale in ways that were never possible
  • Online media is ineffectively regulated. While the internet today is very different from how it was even ten years ago, it is largely self-regulated, and these regulations are unevenly and unfairly distributed
  • Online media producers, like any media producer, are primarily motivated by financial gain
  • Online media can be highly manipulative of vulnerable audiences, for example children and adults with mental health issues
  • Online media is controlled by a small group of incredibly financially successful conglomerates, including Microsoft, Google, Apple and Amazon

Extension questions


  • To what extent does Zoella use Youtube algorithms in order to manipulate her audiences?
  • In what ways does Zoella's content differ from 'old' media products, for example TV shows and films?
  • What are some positive aspects of online media in general and digital convergence in particular? How can marginalised groups, such as those with mental health issues use online media in positive ways?
  • You have the the option to push a button and to irreversibly destroy all digital media. Would you push it? Why?