Netflix is on a mission to get us hooked – and rot our brains
With ‘Shuffle Play’, the service seems intent on keeping us glued to trash TV, says Ed Power
In a year in which everyday life has been transformed into a terrifying but also incredibly dreary dystopia, Netflix has just taken the Orwellian biscuit. The company is trying out a hugely creepy new “shuffle” option, which allows the viewer to sit back and let the streaming service decide what they should watch. The robot uprising is here – and coming for your remote control.
Shuffle Play is currently being rolled out for smart televisions (TVs with the internet built in), though the presumption is that it will eventually be available wherever you access Netflix. And when the company says that it is doing the choosing, what it really means is that its mysterious algorithms are gazing beadily at your viewing habits and reducing them – and you – to a string of data.
Having crunched the numbers, Netflix’s machine intelligence then selects a programme calibrated to chime with your tastes. It’s like the worst imaginable remake of The
Terminator, in which a Teutonic android travels back through time, knocks at your front door – and then insists you sit through Floor Is Lava for five hours straight.
Netflix isn’t alone in embracing the almighty algorithm. BBC’s iPlayer, Amazon Prime Video and YouTube likewise use mathematical-based recommendation. However, Netflix has been in the numbers game far longer than its broadcasting rivals and has built its entire business model around computational breakdowns of subscriber viewing habits.
As early as 2006, when it was still a postal DVD rental company, Netflix offered a $1 million prize to anyone who could create an algorithm for its planned streaming service. Six years later, it acquired its first big original hit,
House of Cards, after analysis identified a holy trinity of subscriber favourites: Kevin Spacey films, movies directed by David Fincher and Andrew Davies’s original House of Cards from 1990, which was regularly featured in lists of the best TV series of all time and was still hugely popular. Who needs commissioning editors – or humans, really – when you have the implacable logic of pure data?
There was something deeply chilling about how Netflix took this information and ran with it. It proceeded d to cut 10 different House of Cards trailers. railers. David Fincher fans saw a promo omo that emphasised his moody camerawork, those who enjoyed television with strong female leads were shown own one that underlined Robin Wright’s starring role. Kevin Spacey devotees got lots of Frank Underwood d gazing into the camera and letting fly with his Evil Bill Clinton impersonation.
And that was just the start. t. Today, your experience of Netflix – in particular, the shows you see e on the home screen – is rooted entirely in your data profile.
Such is Netflix’s dependence on algorithms that it doesn’t even bother to market much of its output. If the numbers think you’ll like a programme, then it will find you. Otherwise, it might as well never exist.
This has created an entire substrata of “ghost TV” – content available to watch but which you’ll never find out about because you don’t meet the algorithmic profile. When musical rom-com was released on Netflix last year, for instance, its creator Josh Safran complained that it seemed to vanish into the digital thin air. “What if you made a show and no one noticed?” he wondered. “I personally feel like it never came out.”
His frustration is understandable. was barely promoted; the media appeared entirely unaware of its existence. The problem was that if Netflix didn’t judge you to be the sort of person who might like a musical rom-com – someone who, for instance, binged on or Mamma – then it didn’t even inform you that Soundtrack had been added. It was sealed off, forever hidden behind the algorithm. “I have my own thoughts on trusting an algorithm instead of reaching out to your audience,” Safran said.
The other obvious complaint is that the algorithm simply isn’t very good. Numbers may not lie but nor are they apparently much cop at randomly picking the best telly to watch. Based on the shows that Netflix recommends to me, it seems to have concluded that I enjoy Formula 1 (no interest), American comedy (maybe, but not really) and Adam Sandler (nope). It’s as if it took a quick gander at my gender and then had a wild guess. Someone out there has to like Adam Sandler, right?
A more serious objection is that allowing Netflix and its competitors to make all our decisions for us cannot be good for anyone. Not for the people who create these shows and whose careers are now dependent on the dark arts of mathematical modelling. And not for the consumer, whose brain has already partially turned to gloop thanks to our addiction to Google and social media.
And, of course, there is also the issue of privacy to consider. Netflix isn’t merely monitoring what we watch – it’s keeping tabs on the how, when, why and length of our viewing too, tracking whether or not we stay with a show to the end, if we press pause and what we watch next.
This information is meant to refine our TV experience even further. And yet that experience is becoming increasingly empty. So much streaming TV feels “samey” – like fast-food. You end up stuffed, but undernourished. But then, maybe the goal isn’t to sate our desire for entertainment. Perhaps it’s the opposite: to keep us watching. Always looking, never finding something that quite hits the spot.
“The problem is that the AI isn’t built to help you get what you want,” ex-Google engineer Guillaume Chaslot told a technology news website last year. He was talking about YouTube, but the lesson applies equally to Netflix. “It’s built to get you addicted to YouTube. Recommendations were designed to waste your time.”
The algorithm has decided that I like Formula 1, US comedy and Adam Sandler – I don’t