Being ill during the holidays can have its benefits. I got to watch A TON of movies that I had missed throughout the past year. One of the films I saw was HER by Spike Jonze and it blew me away! So much so that I immediately jumped online to read through all of the articles written about it in the past year. In the process, I also found an amazing blog AND book on SciFi interfaces and what interaction designers can learn from them. But let’s start at the beginning.
One of the appealing aspects of the movie was not so much that futuristic interaction design was played a large role in the story. In fact, the interaction design displayed had nothing on the eye candy shown in other movies of the same year e.g. Guardians of the Galaxy. But the movie left me with a few interesting questions that I wanted to do some research on.
– Why does the relationship between Samantha (an AI) and Theodore work so well for the viewer?
– Will technology ever be as beautifully invisible and ubiquitous as it is in HER? And if so, what would the challenges look like?
As I might have been the last person on earth to see the movie I assume that whoever reads this has seen the film. So I won’t be going into too many details of the story and might spoil aspects of the story for you. You have been warned!
Nostalgic near future technology
Look Ma! No hands!
One of the amazing accomplishments of the movie is that it manages to show us a “near future” with highly developed technology well integrated into everyday life. Therefore we can relate to the given reality more than to the worlds of other movies in which society has come this far technologically speaking. This is not only because of the absence of typical fantastic Sci-Fi story elements such as aliens or time travel but also because the technology shown is nearly invisible, somewhere in the background, reacting to human needs in just the right moments. This is very different from the semi-transparent glow fests that occupy entire rooms in other movies of this genre (yes, I was thinking of Avatar). The human-machine touch points are beautifully solved, gentle and nearly poetic.
In the opening scene, we see a closeup of Theodore, who seems to be talking to a loved one, speaking of his love for this person. When we zoom out we see that he sitting at the desk at his job (beautifulhandwrittenletters.com) and is “writing a love letter” for one of his clients. The computer turns Theodore’s words into handwritten letters presumably in the handwriting of the clients, which are mailed off at the end of the day. For all we know this happens without the touch of a keyboard or a single push of a button. There is no device barrier between human and computer. The interaction is solved with a voice interface.
When Theodore leaves work we are introduced to his ear device and a device that looks like a vintage cigarette case which is able to display visuals and has a built-in camera. Theodore again navigates through his messages and feeds of the day through a voice interface, earphone and the handheld visual device. The voice of the OS is robotic, stoic and emotionless.
What Theodore is doing in these first scenes is highly relatable to us. People with desk jobs might have a very similarly structured workday and commute. Theodore, however, manages to do these tasks without his fingers glued to a keyboard, a mouse or touchpad, a concentrated glance at various screens. Even though it is exactly the same things we do today, Theodore’s version seems more natural to us, reminds of interaction with humans before we had the gadget barrier and almost feels nostalgic.
Theodore and Samantha: a human-AI relationship
OS1 – “It’s not an operating system, it’s a consciousness”.
When Theodore updates his personal OS to OS1, Samantha (Theodore’s version of OS1 and soon to become love interest) is introduced. Samantha also works with a voice interface. Her voice, however, is not mechanical, monotone or robotic. Instead, she follows human social conventions (“Make it so: Interaction Design Lessons from Science Fiction” by Nathan Shedroff, Christopher Noessel: chapter 6 Sonic Interfaces / Voice Interfaces). When Samantha is first activated her first question is “How are you doing”. Samantha even gives herself her name that from then on Theodore uses when referring to her. Samantha is intuitive, capable of learning and adapting to the needs of her owner and as we find out later, even autonomous. When she leaves messages for Theodore the notice is displayed on Theodore’s handheld visual device in “her handwriting”. In short: she seems and sounds human. Sam’s voice has the same resonance and body as all human voices in the story. She sounds as if she were in a room with Theodore rather than having her voice transmitted over a crummy mono earpiece-speaker.
The whole human-computer interaction was very carefully constructed to make us, as well as Theodore anthropomorphizes Sam, which is a typical human-user act (“Make it so: Interaction Design Lessons from Science Fiction” by Nathan Shedroff, Christopher Noessel: chapter 9 Anthropomorphism/ Humanness is Transferable to Nonhuman Systems). The general production design further supports this through its vintage design choices giving this future a warmly feel in which it is very easy to forget that Sam is an AI. (My personal favourite design choice: a safety pin holds the handheld device in place in Theodore’s shirt pocket so Sam “can see” when they are on trips together. This solution is refreshingly analogue).
Another reason why the relationship between Sam and Theodore works so well is the power of human imagination, which is also addressed multiple times in the story. During the telephone-sex scene, Theodore imagines (and the viewer sees) a nude starlet he had viewed pictures from earlier that day. This fantasy is interrupted when his partner suddenly asks Theodore to “choke her with the dead cat” – a fantasy for his partner but a turnoff for Theodore. Another scene is located towards the end, when Theodore is imagining himself hugging Sam in the snow-covered mountains where they had spent their vacation together. In this scene, Sam has dark hair, which might be one of the reasons why the date with the human surrogate lover, that had been arranged by Sam (earlier in the movie) went so terribly wrong. The surrogate had blond hair.
An additional reason why this experiment failed can be found in Chris Noessel’s keynote:
Even though OS1 / Samantha is clearly “near-future” technology the reason why the relationship works so well for the viewer is because of human nature. When technology is as enhanced as OS1 and barriers that we have today are broken down the reactions become human-natured and follow different patterns as predicted by user stories. In Theodore’s case, we see anthropomorphism, love, heartsickness. Even though Sam starts to reveal/find her intellectual capacities and superiority and the whole relationships head towards an uncanny valley Theodore is left with a broken heart at the end of the movie.
The future of ubiquitous technology and the challenges
The next logical question is if a story and technology like in HER were possible in the real world. In his keynote, Christopher Noessel asks the question who would build an OS such as OS1 (article on scifiinterfaces.wordpress.com). His conclusion is military, academia or corporation. He further assumes that a corporation would be most likely to release a product such as OS1. He however does not believe that a product as OS1 would ever be introduced to the consumer market. One reason being that the sheer fact that the OS can fall in love with its owner is too unprofessional, meaning the system was too buggy for a release.
I do not agree with this assumption because I do not believe that a human product/ interaction designer can foresee all possible user or AI stories (OS1 is adaptable and is constantly learning) and their outcome in a system mimicking humans. The reason for the unhappy ending of Theodore and Sam’s relationship is the “malfunction” of Sam falling in love and encouraging Theodore to do the same before realizing that Theodore, the human is not enough of a partner for an AI. The designers hadn’t foreseen these possibilities.
I don’t think this would keep corporations from releasing a product such as OS1, though. Just take a look at all the products out there today that do not have a thorough concept, design, production and QA. I am pretty sure someone is currently trying to build OS1 somewhere in this world and shiver because of the Franksteinian monster visions I get just by thinking about that.
In the episode Martha loses her partner Ash in an accident. A friend signs her up for a new service that allows people to communicate with their deceased loved ones. This is made possible because the service creates an AI that is fed with public data left behind by the deceased in social media and the general internet. After a slow start that is packed with skepticism on Martha’s behalf (she does not want to get attached to something digital and not real), she embraces the new possibility and spends quite some time talking to Ash. However Ash then suggests Martha to take the next step and book an additional, experimental service of the company. The additional service is a very own Ash robot (The technology is not specified.), that is shipped in a box, inflated and brought to life in a bathtub filled with water. This is when Martha starts to see the flaws in this AI because robot Ash knows things the real Ash would not know (as he can learn from online resources) and reacts in ways the real Ash would not react (as the information on the reaction is not available in an online database). The relationship turns sour and ends with Martha keeping robot Ash in the attic, not being able to abandon him completely.
Both stories are love stories that do not end well due to “malfunction”. In “HER” OS1 was not even intended to fall in love with it’s client, while in “Be right back” the incarnated AI is not human enough. This is because it is impossible to predict the use cases for products with such depth and possibilities, products that are endlessly scalable, adaptable and self adjusting.
I personally see the challenges of ubiquitous computing in holding the balance between predictable or orchestrated functionality and companionship. I am not sure if this could go as far as the creation of an AI, without risking various scenarios in which the functionality could grow unintentionally and lead to “malfunction”. It can however surely lead to more than the single task IoT objects that we have today.
Even if we know the math it does not mean that we can compute social psychology. But I’m pretty sure we’ll try anyway.
Links and research resources:
– When done reading this next article on designworklife.com also check out the informative Reddit AMA Spike Jonze did.
– If you are interested in this subject you should defiantly read “Make It So: Interaction Design Lessons from Science Fiction” by Nathan Shedroff, Christopher Noessel . Also check out their amazing blog www.scifiinterfaces.com
– My most utterly respect goes to K.K. Barrett for the marvelous production design, as well as Geoff McFetridge, who was responsible for the interface design. Chapeau! Also check out the interviews on their work on HER with Design & Architecture: K.K. Barrett’s interview, Geoff McFetridge’s interview
– Print Mag wrote some more on Technology and Love.
Websites for further research:
– fakeui.tumblr.com has gathered some pretty good examples for inspiration.
– Another promising source might be KIT FUI: “an IMDb-like database that makes it easy to research screenshots, videos and the designers of these FUIs.” (PC only, I don’t have a PC so I can’t vouch for it.)
All images are taken from www.herthemovie.com