- Vision: ‘Resurrecting’ the dead from their social media history
- Type: Popular culture – TV
- As Seen In: Black Mirror episode ‘Be Right Back’
- Envisioned: 2013
- Visionaries: Charlie Brooker
- Target Date: Unknown – though the subtle advances in cars and computers shown in the episode suggest it’s near-future
Black Mirror is writer Charlie Brooker’s very modern take on The Twilight Zone, an anthology show that revels in showing us a dark side of the technology that’s become ubiquitous in our lives. One of the big reasons for the show’s international success is its ability to present a recognisable version of the modern world, while pushing the sci-fi elements just enough to ask extremely pertinent ‘what if?’ questions. Never has this been more true than in series two opener ‘Be Right Back’.
In the episode, Martha (Hayley Atwell) suddenly loses her partner, Ash (Domhnall Gleeson), in a traffic accident, and finds solace in software that uses social media histories to impersonate the deceased. Quickly moving on from text conversations with an Ash-themed chatbot, Martha starts talking on the phone with a spookily accurate recreation of her late boyfriend. Then, when she’s well and truly hooked, she signs up for the deluxe package: a live-in android with Ash’s personality.
Of course, it wouldn’t be Black Mirror if there wasn’t a catch, and Martha quickly realises that this idealised automaton lacks the human flaws that made Ash who he was. But while mail order androids remain well and truly in realms of science fiction, much of ‘Be Right Back’ is proving spookily prescient in 2021.
- Hardware: The early stages of the ‘Be Right Back’ experience make use of widely available hardware – laptop computers and smartphones.
- The experience starts with an email claiming to be from a loved one who’s passed away. A red button on screen prompts you to ‘Touch to Talk’, and opens a chat function where you can talk with an AI replica of the deceased.
- The system draws data from the deceased’s social media history to make the responses as convincing as possible.
- After an upgrade, the system uses recordings mined from social media sites to create a convincing simulation of the deceased’s voice. It then uses voice recognition software to interact with the user in real-time conversations.
- The data exists in the cloud, so can be used across multiple devices.
- The AI simulations are constantly connected to the internet, so can check facts instantaneously. This functionality can be turned off if the user finds it disconcerting to talk with someone who knows everything.
- The most advanced “kind of experimental” package sees the AI simulation uploaded to a walking, talking android – an almost exact likeness of the deceased.
- Androids are an approximation of the original human, based on images posted online. They’re freakishly smooth and – because most people usually upload the more flattering photos of themselves to social media – have a tendency to look like the deceased did “on a good day”.
- Androids are anatomically correct and have the ability to make subtle adjustments to their appearance. If, for example, a mole isn’t where it should be, the unit can make one appear on their skin.
- Experience: Using sophisticated AI algorithms, the system mines social media to respond to the user in a manner practically indistinguishable from the real thing.
- The system can create more convincing simulations of “heavy users” of social media. This is simply because there’s more information available online.
- Online chats and phone conversations make the user feel they’re having a conversation with the real person. The voice-only version of the tech is reassuringly convincing.
- The software appears to be programmed for upselling. Once it’s got you hooked on the experience, it drops in a subtle hint that a more sophisticated version is available – at substantial financial cost.
- The ‘Be Right Back’ androids arrive in a dehydrated form via mail order. The user needs to place them in a bath – along with the nutrients and electrolytes provided– to activate them.
- The androids move and react like real people, and can simulate eating and drinking – even though they have no need for sustenance.
- They can also perform in the bedroom, but not in the same way as the deceased. Because they (probably) never recorded such information on social media, the unit has no idea how they’d behave in intimate situations. It instead uses pornographic material to inform its technique.
- The android has no need for sleep and, thanks to a major programming flaw, doesn’t close its eyes or ‘breathe’ at night. This is extremely disconcerting for the user.
- The android also struggles to read social cues, and will blindly follow the user’s commands. It won’t argue back as a real person would.
- The android seems to accept that it will never be a replacement for the genuine article.
- The android is unable to travel more than 25 metres from its activation point, unless accompanied by its user. In other words, once you’ve activated it, you’re stuck with it – they’re even resistant to jumping off cliffs.
- Company: The name of the company who manufactures and supplies the technology is never mentioned in the episode. Robo-Ash even arrives in an unmarked box.
- Economics: The exact price of the technology is never mentioned on screen. It’s clear, however, that there is a cost every time you upgrade to the next level. The android package is – Ash admits – “not cheap”.
- IP: The system appears to grab data from any and all social media platforms available.
State of Play (December 2021)
- Hardware: Speech recognition-based software like Siri, Alexa and Cortana is now a fixture in our lives.
- Artificial intelligence technology is also improving all the time – while it’s not genuinely self-aware or intelligent, it has the ability to synthesise more and more convincing responses to our questions.
- So-called ‘deepfake’ technology exists that can generate spookily accurate recreations of faces and voices. The full extent of the tech’s capabilities – for good and bad – is yet to be realised.
- It is, however, becoming more widespread in Hollywood. When a young Luke Skywalker cameoed in Star Wars TV show The Mandalorian, his voice was completely synthesised using voice cloning software called Respeecher. The show’s sound effects team fed archive recordings of a 20-/30-something Mark Hamill into the computer, which then used them to fabricate a convincing voice performance. The present-day Hamill’s voice didn’t feature at all.
- Despite numerous scientists working in the field of cybernetics, creating a realistic android that can move convincingly enough to pass as human remains elusive. The so-called ‘Uncanny Valley’ is not our friend in that regard.
- Experience: Just a decade ago it would have been unthinkable to find ourselves regularly talking to machines, but these days we ask questions of Siri, Alexa and Cortana without giving it a second thought.
- We frequently interact with customer service ‘chatbots’ online, which give the illusion of talking to a real-life human. They are, however, unable to deal with more complex requests.
- AI can also now write its own music and even newspaper articles – like this example published by the Guardian.
- Following the death of her friend, Roman Mazurenko, Eugenia Kuyda (founder of AI start-up Luka) created a digital avatar as a memorial. Available as an app, the bot would reply in the style of Roman. You can read more at the Verge.
- Economics: There would undoubtedly be a huge market for anything that gave the bereaved the chance to ‘talk’ to their loved ones – whether the experience is real or not. Indeed, Black Mirror producer Annabel Jones has said, “In some respects it’s the modern equivalent of a medium.” Brooker, meanwhile, told Wired that in early versions of the script, “There was going to be a beat where you realise what a money-making scheme it is. There was a point where [Martha] runs out of credit and has to top it up.” Of course, the ethics of exploiting grief for financial gain are another matter entirely…
Reality check: extremely plausible, except for the android bit
Google and social media platforms collect so much of our data that they already know more about us than our own families and friends. Using this information to create a facsimile of ourselves after our deaths seems frighteningly plausible. Increasingly sophisticated deepfake technology also means that eerily realistic video and audio simulations of human beings are already with us.
And those mail-order, boil-in-the-bath androids? Thankfully, they’re still some way off…