A transcript of my presentation for Sci Gallery Melbourne, Friday 17 August 2018, as part of National Science Week. It was an introduction prior to screening the film ‘Ex Machina’ (2014) at Federation Hall, The Victorian College of the Arts, Melbourne.
“Before you settle in to watch the film, I want to get the cogs turning a little around major themes- no spoilers! If you haven’t seen the film before, don’t worry.
When I was a smaller person, my Dad made the somewhat naeive error of insisting that I learn that game of supreme tactics, operating somewhere in between excruciating deliberations and then spurts of careless bravado- the game of chess, of course!
He quickly overcame his enthusiasm when I started to win, and wanted to deflect my appetite for conquest onto another, perpetually willing competitor. So we loaded up the old Windows XP computer with a CD-ROM of a chess program- he was off the hook.
I had a head full of imaginings around the tale of Deep Blue VS Gary Kasparov, the infamous tournament loss of a human chess champion against a computer. Even so, I made sure I dialled the difficulty settings all the way back to ‘easy’. Now this was to be my first experience of friction against a perceived algorithmic injustice, an encounter of disparity between human and non-human action taken towards a common objective- to win the game. Time and time again, as I played, rather permit my win, the chess program would enforce the second optimal result: flashing up stalemate- stalemate- stalemate! I remember feeling outraged at the unfairness of it all, at what seemed like a cold ungraciousness in what I imagined to be my opponent’s persona.
Dad reminded me that was a fairly simple case, though, of placing blame on the puppet masters- the developers who had written the chess program’s code. They were responsible for the unyielding, relentless tendencies of their creation. The actions of this program, hardly representing an autonomous thinker, was only a proxy for the whims of its makers.
This brings us to the film we’re here to watch today- Ex Machina. As you might suspect, the story doesn’t offer up such a simple case. This movie is actually one of my favourites in a fairly recent crop of thoughtful sci-fis. Beyond the fanciful, glossy elements of typical sci fi- ultra advanced tech, cyberpunk urban sprawl, and very tight jumpsuits- audiences are encouraged to ask big picture questions of judgement, empathy, conscience beyond the human. These are what’s termed ‘near future’ concerns- not too far away that we’re talking about transcending our bodies and melding into the universal hive-mind, but along the lines of social media becoming a bit more all consuming, our phones doing more for us than ever, or the internet of things being in pretty much in everything. The hardware we already have in our pockets and the software we subscribe to with full consent, in the near future. In Ex Machina, there’s three relevant concerns around the future of tech that particularly struck me, and maybe you can think about these as you watch tonight.
The first arises from that story of stalemate in my chess adventures, and the agency to take certain actions. Where does that distribution of responsibility fall when there are actions of intelligence made, rather than born? How can these actions be judged if there are involuntary, prescribed differences in someone’s moral compass and ability to behave? Is it just enough to create a faithful, perfect copy of our own psyches, suspended in wetware gel? Or should we take our tech obsession with upgrade and improvement to engineering a better version of us- the next unmissable new release, a host of exciting new features, a symbol of advancement beyond perfect. Machine learning researchers surveyed by a specialist unit at Oxford University answer with some conviction. They predict AI will outperform humans in many activities in the next ten years, such as translating languages (by 2024), driving a truck (by 2027), working in retail (by 2031) or working as a surgeon (by 2053). These researchers believe there is a 50% chance of AI outperforming humans in all tasks in 45 years and of automating all human jobs in 120 years. (1)
The title of the film, as some are probably aware, relates to the phrase : ‘Deus ex machina’. The term was used in ancient Greek theatre. It meant, ‘God from the machine’, where an intervening character would fly onto the scene, suspended high above the stage via a mechanical contraption- the machine. They would change the course of the story through powers often unexpected, or supernatural. In the case of tonight’s film, the title itself confronts us with the notion of the machine unencumbered, disconnected from a deity, master or overseer. Thus given elevated levels of autonomy, we are invited to consider an experience of sensation and action, indeed, a consciousness, as it might occur ‘ex machina’- from the machine… alone.
Now I recall a few occasions of being in the car with my partner, who assures me that he is a master of navigation, a genuine wayfinder in a world of directionless chaos. Inevitably of course, we require a helping hand- courtesy of Google Maps. He will protest: “Oh- shut her up! She’s always wrong anyway, she doesn’t know these streets like I do.” Google’s voice navigation is set by default as a female voice, and it’s not a big stretch to imagine a extended persona that goes along with this- one that is specifically gendered. Sure, it’s just a voice- but like Scarlett Johannsen’s character in the film Her, sometimes that disembodied voice might be all it takes to embed itself into cultural norms and expectations. If we tap into the archives of tech media cultures, from sci fi gaming, film, graphic novels, and the like, we encounter an entrenched history of female gendered virtual assistants.
I grew up console gaming with a character called Cortana, an A.I helper in the Halo franchise. Now, a virtual assistant called Cortana issues cheery pronouncements on the weather from the desktop of my Windows PC, with an twangy American woman’s voice. My new Samsung phone rather unfortunately pits an A.I assistant called Bixby- Samsung proprietary-against the cohabiting, also female voiced, Google Assistant. Both carefully avoid mentions of their Apple counterpart Siri, attending to such menial tasks as turning my Wi Fi on and off, or calling someone handsfree. These current day assistants also orchestrate the latest in smart home technology, bonafide domestic goddesses that are not yet seen- only heard.
Where the rub lies is the continual reinforcement of these artificial, abstracted, but identifiably female subjects into a deepening history of servitude and unquestioning compliance. This not only problematic for it mirroring retrograde gendered arrangements of imbalance and inequality in humans, but it is establishing such a precedent in nonhumans- the sort of thing from which dystopias are dreamed. Much of the control and initiative to shape artificial service assistants or ambient intelligence lies with tech workers, a traditionally male held domain. This is tied up with the tendency for contemporary society to applaud, glamorise and bestow incredible riches, cultural impact and power upon the tech heroes that design these systems. Gender politics are already at play in how we think, speak about and regard service technologies today. This can hardly get simpler as each iteration brings advanced complexities to the fore.
The final point I want to bring up around artificial intelligence is more currently tangible, gaining greater exposure with every scandal that hits the headlines. This is the surveillance, algorithmic profiling and hijackable features that are embedded in, or accompany, the technology we use today. I have a lot of slightly paranoid discussions about how my internet searches will suddenly create relentless advertising, despite my interest in dog beds or toilet seats being largely one-off. I’m not a toilet seat connoisseur or collector, you know- can’t we go back to pop up ads that fill up the screen, strobe crazy colours and say I’m today’s lucky winner??!
Tech ethics researcher Katja De Vries argues that the way in which we experience ourselves necessarily goes through a moment of technical mediation- the apparatus we use, those adornments and additions to ourselves in a physical sense and the way in which an online portrayal of ourselves defines a sense of who we are. She argues that there is growing societal concern around the impact of algorithmic, computational profiling on our sense of identity. (2) This is most obviously occurring online, in which algorithms thrive on reconfiguration of human identity, claiming alleged infatuation with toilet seats at the flimsiest of diagnoses.
This type of thing poses a sense of existential questioning around who or what influences or manufactures a sense of self and even opens up the possibility of preference-based discrimination through increasingly detailed, cross referenced and nuanced examinations of our behaviour through artificial intelligence. With a massive source of information from which to compare us among many, many others like and unlike us- the internet- the way in which algorithms capture us is not a perfect reflection. Rather, a new twist is generated in our stories, and who we appear to be from an outside point of view.
That all being said- I leave you now to enjoy the film!”
- Grace, Katja, John Salvatier, Allan Dafoe, Baobao Zhang, and Owain Evans. ‘When Will AI Exceed Human Performance? Evidence from AI Experts’. ArXiv:1705.08807 [Cs], 24 May 2017. http://arxiv.org/abs/1705.08807
- Vries, Katja de. ‘Identity, Profiling Algorithms and a World of Ambient Intelligence’. Ethics and Information Technology 12, no. 1 (1 March 2010): 71–85. https://doi.org/10.1007/s10676-009-9215-9.
Image source: https://melbourne.sciencegallery.com/sci-film-ex-machina