For most Digit readers, HAL 9000 was probably the first interactions they had with an Artificial Intelligence that acted as an assistant. If you Gen-Z newcomers are confused, HAL 9000 is the infamous sentient AI that acted as the assistant on a spaceship in Arthur C. Clarke’s 2001: A Space Odyssey. This, possibly, is the most oblique way to get into the nitty gritty of the AI voice assistant that we know as Siri. Nonetheless, it should be known that one of the earliest projects to create a voice assistant was known as ‘HAL’, and had the tagline “HAL’s back – but this time he’s good.”
Soon, Defense Advanced Research Projects Agency (more commonly known as DARPA) decided to take an interest in this. This is the same DARPA that is known for inventing the earliest form of the Internet as well as modern day GPS. They decided to try to create a voice assistant that would assist soldiers in real-time.
Siri begins: The origin story
Siri was founded in 2007, but the research it was based on came from a DARPA programme that started way back in 2003. They called it Cognitive Assistant that Learns and Organizes (CALO). Acronyms have an odd habit of being derived from the coolest of origin languages, and CALO isn’t one to break away from said habit. Derived from the Latin word ‘calonis’ or Soldier’s Assistant, CALO was the foundation on which present-day Siri is built.
The story of Apple’s acquisition of Siri is in keeping with the rather covert nature of things up until then. In February 2010, about 3 weeks after it first appeared on the iTunes store, Dag Kittaus (Co-founder of Siri) got a phone call from a mystery number asking to meet – it was Steve Jobs. In a later interview, he is quoted as saying, “Of course it was a great moment when Steve Jobs called and wanted to buy my company. It was surreal. When I heard that it was him, I knew we had made it big. In advance, we were pretty confident that the technology we had developed was so startling that we would get some kind of breakthrough. Steve was the first caller.”
Rather strangely, a lot of what made Siri such an exciting purchase for Apple wasn’t implemented immediately. Apple released a beta in October 2011 that was fully integrated with the iPhone 4S. Shortly afterward, it was removed as a standalone app from the App store.
From promise to purpose: The problems with becoming a Big Shot
Apple faced many teething problems while integrating Siri into the iPhone. Originally, Siri was basically the centerpoint of a database of over 40 services and was programmed to understand the user query and find the best combination of results. Even as Apple was modifying Siri to broaden its horizons, it was diminishing its returns. Take for example, the restaurant booking feature that was available on the original Siri way back in 2010. While it doesn’t seem like that big a deal right now, it was one of Siri’s biggest unique selling points. It wasn’t offered to Apple users till late 2012, almost two years after it was officially released.
It was theorised that this is because of Apple’s acquisition of Siri. When it was a burgeoning voice assistant application, there was little-to-no red tape involved when making associations with industry leaders in fields such as Search Engines, Restaurant and Hotel Booking, Location Detecting et al. However, with Apple now owning Siri, making such associations involved a lot of negotiating and even more paperwork.
Adding to this was the relative infancy of the entire field of Natural Language Processing, which meant further delays in rolling out the best version of Siri. As a corporate entity, Apple had responsibilities towards its customers to ensure that Siri functioned well in the nearly 100 countries that Apple was selling its products in. The iPhone was expected to recognise thousands upon thousands of unique accents in order for Siri to be rolled out. In the process, they ended up making a decent application rather than the ground-breaking development they promised in the 2011 keynote address.
Siri was meant to be a cross-industry genius instead of the single-use robots that preceded it (read: Microsoft Office’s tragic Clippy), and all it needed was access to a multitude of unique Application Programming Interfaces (or APIs) that companies often gave out to 3rd parties. It was meant to be paired with potentially thousands of APIs to supply a continuous stream of information to its user. Not stopping at that, there were dreams to understand the user’s idiosyncratic way of talking and respond in kind. Such dreams have remained in the pipeline for too long, but Apple’s Siri isn’t without promise.
A nicer HAL: Siri becoming a part and parcel of every Apple device
Despite all the idealistic criticism it faces, Siri has become an integral part of every single Apple device. It has reached a point where it is almost unimaginable to hypothesize the future of Apple without simultaneously involving Siri at every turn. It fulfils a multitude of daily tasks – calling, texting, emailing, setting alarms and reminders, checking the weather, and even doing maths problems. It can recognise inputs involving native apps like Twitter and Facebook, as well as search for documents in different folders of the device. It also has the ability to navigate through traffic, find the nearest whatever-you’re-looking-for, and keep you up to date on the rise (or fall) of your favourite stocks.
Beginning with iOS 10, Siri finally started to interact with third party apps like WhatsApp, Skype, PayPal, and Uber. However, by far her most recognisable calling card has to be her dry sense of humour. Before Apple white washed its language, Siri was capable of using cuss words with an alarming regularity. The current version can tell you a joke, plunge you into the absurd depths of an existential crisis, insult your best friend, or call you a loser – all at sound of your voice.
10 years ago, conjuring up an AI voice assistant capable of all this would’ve taken a figurative leap of imagination. Now, we’re at the cusp of a revolutionary change.
It’s an AI eat AI world: Siri’s many competitors
With the incredible success of Siri, tech giants everywhere have started excavating CALO to model their own voice assistants. Google has Google Now (or Assistant), Microsoft has Cortana, and Amazon has Alexa. We are staring down a future where technology understands us much better than we understand technology. Shawn Carolan (a Siri investor) says “Take everything you do in a day and just condense it down from 15 minutes down to 30 seconds. You can just express your intent, and it gets done. You just became 30 times more powerful human.” Dug Kittaus adds,“The future of virtual personal assistants is to make it so we don’t have to think so much and work so hard to do things that are possible. It’s less about survival and more about exploring the world.”
Such leaps and bounds in technological conventions were met with societal conventions. Upon being asked questions about certain controversial issues, Siri has had very politically incorrect answers.
Navigating the waters of the politically correct: Siri’s Odyssey
It is unfair to say that this is a Siri-centric problem. Voice Assistants everywhere have had controversial issues turn up on their doorstep every now and then. In 2015, when asked the definition of the word b**ch, both Google Now and Siri replied that it was ‘black slang’. Earlier, Google Now had mistakenly tagged a photo of an African-American couple as ‘gorillas’. Both were rightly chastised. Siri also had a protracted controversy about women’s hygiene and abortion. The NYT later revealed that it was more to do with Siri’s lack of understanding than with Apple’s stance on women’s issues. However, new stuff keeps popping up. For instance, apparently Siri can find the escort services closest to you, can locate the nearest marijuana store, and can help you dispose off a body but can’t understand the terms “IUD’s” or “contraceptives”.
The word ‘inevitable’ should just be repeated over and over when such issues come to light, as it is a self-taught program that is trying to navigate the complex geopolitical issues that even human beings can’t fully wrap their heads around.
The Ultimate Question
In 2001: A Space Odyssey, HAL 9000 tries to takes some bad decisions based on what it believes is the greater good. The countless Terminator movies talk about Skynet – a systematic overhaul of society by machines – leading to an apocalypse. Pop culture is full of references to an AI-based doomsday, yet the closest parallel to Siri is in Spike Jonze’s Her, where the protagonist ends up falling for his customised voice assistant. The voice assistant is very much like an advanced Siri with her subtle wit and deeply human empathy. The film investigates our ability to find emotion in a complex organism, despite knowing that it is made up of 0s and 1s. Perhaps the ultimate voice assistant will be the one that allows the user to fall in love with it.
Or perhaps it is another technological rabbit-hole, a bubble that will burst as soon as the next big thing surfaces. Who can tell, really. Perhaps 50 years from now, if we use voice assistants with the same regularity that we use instant messaging apps in the present, we will ascertain Siri to be the start of something special. Up until then though, we will have to wait and see.
This article was first published in the September 2017 issue of Fast Track which comes with Digit magazine. To read Digit’s articles first, subscribe here or download the Digit app for Android and iOS. You could also buy Digit’s previous issues here.