Beginning this year, I am seeing more frequent mention of artificial intelligence (AI)-driven, conversational user interface (UI) in technology media and among my technology-oriented friends. This is no surprise, AI and conversation (either verbal or written) based UI has been around for a while. Siri, Cortana, OK Google are familiar features. Amazon Echo, Magic (app), Slack, WeChat, Facebook's M are more recent, high profile additions to the category. The purpose of this writing exercise is to explore how and why this new trend is flourishing, and whether they will evolve into something significant.
Conversational UI will undermine visual interface design, according to
Matty Mariansky:
But, let me say it again. The entire field of visual interface design—everything we know about placing controls, handling mouse and touch interaction, even picking colors—will be affected by the switch to conversational form, or will go away altogether. Store that in your brain’s temp folder for a little while, then take a deep breath. Let’s move on.
Text-based, or messaging interface is still what people use the most to communicate, according to
Sarah Guo:
If we take college students as a leading indicator, I’m not alone. In 2014, college students (at least in this interesting study at Baylor) spent ~9 hours a day on their smartphones. Of those 9 hours, almost 30% of that time (2.4 hours) was spent either texting or emailing. Messaging has become the dominant digital experience for many young people.
Sarah Guo also brought a great point on the platform advantage of popular messaging apps:
the popular messaging clients (mostly FB and WeChat, potentially Slack) have been smart enough to develop ecosystems that trade their distribution power and real estate on your phone for third party invented features, behaviors and content that, in turn, increase their platform’s relevance and stickiness.
Text-based interface is praised for its simplicity, according to
Jonathan Libov:
In contrast to a GUI that defines rules for each interaction — rules which, frustratingly, change from app to app — text-based, conversational interactions are liberating in their familiarity. There's only really only one way to skin this cat: The text I type is displayed on the right, the text someone else typed is on the left, and there's an input field on bottom for me to compose a message.
Text is powerful, according to
greydon 2:
Text is the most socially useful communication technology. It works well in 1:1, 1:N, and M:N modes. It can be indexed and searched efficiently, even by hand. It can be translated. It can be produced and consumed at variable speeds. It is asynchronous. It can be compared, diffed, clustered, corrected, summarized and filtered algorithmically. It permits multiparty editing. It permits branching conversations, lurking, annotation, quoting, reviewing, summarizing, structured responses, exegesis, even fan fic. The breadth, scale and depth of ways people use text is unmatched by anything.
Messaging could change the way we interact with computers, according to
Jonathan Libov:
Messaging is the only interface in which the machine communicates with you much the same as the way you communicate with it. If some of the trends outlined in this post pervade, it would mark a qualitative shift in how we interact with computers. Whereas computer interaction to date has largely been about discrete, deliberate events — typing in the command line, clicking on files, clicking on hyperlinks, tapping on icons — a shift to messaging- or conversational-based UI's and implicit hyperlinks would make computer interaction far more fluid and natural.
It is unknown if those "
invisible apps" can be powered by pure AI, can be scale up at reasonable cost, or can provide good enough user experience without human intervention.
Siri is supposed to be a "do engine", according to its
original founders:
As conceived by its creators, Siri was supposed to be a "do engine," something that would allow people to hold conversations with the Internet. While a search engine used stilted keywords to create lists of links, a do engine could carry a conversation, then decide and act. Had one too many drinks? The ability to coordinate a Google search for a ride home might elude you, but a do engine could translate a muttered, "I'm drunk take me home," into a command to send a car service to your location. The startup's goal was not to build a better search engine, but to pioneer an entirely new paradigm for accessing the Internet, one that would let artificially intelligent agents summon the answers people needed, rather than pull relevant resources for humans to consult on their own. If the search engine defined the second generation of the web, Siri's co-founders were confident the do engine would define the third.
The technological backdrop for Siri was mature in 2007, according to
Huffington Post:
Yet if ever there were a right place and a right time for virtual assistants, the fall of 2007 appeared to be it. Faster wireless speeds, better speech recognition, the rise of cloud computing, the debut of Apple’s iPhone and a flood of new web services made virtual helpers seem attainable at last.
Siri can tap into as many sources of information as third party APIs allow, according to
Huffington Post:
To pull a list of cafés, Siri would tap into data it had organized from over 40 web services that operated like remote, diffuse lobes of its brain. While previous virtual assistants functioned through deep training in a single specialty, Siri had been built as a cross-industry savant with expertise in anything from books to bagels; it just needed access to the application programming interfaces, or APIs, many web companies offer to third parties.
Early on, the Siri developers saw virtually no limit to the routine transactions the assistant could automate. They envisioned Siri’s architecture allowing for any web service with an API — potentially hundreds of thousands of them — to add its database to the do engine.
Siri almost became a selling point for Android phones, due to Verizon, according to
HP:
Verizon thought so, too. In the fall of 2009, several months before Apple approached Siri, Verizon had signed a deal with the startup to make Siri a default app on all Android phones set to launch in the new year. When Apple swooped in to buy Siri, it insisted on making the assistant exclusive to Apple devices, and nixed the Verizon deal. In the process, it narrowly avoided seeing Siri become a selling point for smartphones powered by its biggest rival, Google. (Somewhere in the vaults of the wireless giant, there are unreleased commercials touting Siri as an Android add-on.)
Siri was left an orphan at Apple, which was probably why it seemed to have under-delivered so many years after the acquisition, comparing to what it was capable of when it was an app, according to
HP:
But corporate politics have been unkind to Siri, and the endeavor’s prospects may be jeopardized by its loss of many powerful advocates within Apple. Though Saffo, the Stanford professor and futurist, cautions deciphering Apple’s inner workings is like “trying to understand North Korea,” he ventures that Siri has become “an artificially-intelligent orphan” within the company.
Only one of Siri’s three co-founders, Tom Gruber, remains at the company. Kittlaus left three weeks after Apple re-launched Siri in 2011, and Cheyer quit a year later. Apple’s Forstall, who introduced Siri at its first keynote and oversaw the company’s iOS software, was fired last year. Steve Jobs died the day after Siri debuted. And Luc Julia, who replaced Kittlaus as head of Siri, lasted just 10 months at Apple before leaving in 2012.
Who knew Samsung also had a voice assistant like Siri, called "S-Voice", according to
HP:
More than half a dozen Siri-like services launched in 2012 alone. Samsung debuted S-Voice, a voice-controlled assistant. Nuance, a provider of speech recognition software, released a “Siri for apps” called Nina. Startups Evi and Maluuba each released virtual assistant apps. IBM is working on adapting its supercomputer Watson into a turbo-charged Siri that can help physicians, farmers, Wall Street traders and high-schoolers. And Google has followed Siri with its own conversational assistant, Google Now.
The slow, and relatively closed-door evolution of Siri might be leaving room for its competitors (platform players) to catch up. For example, Facebook acquired Wit.ai in Jan. 2015. Wit.ai is an open and free natural language processing API. According to
Tech Crunch:
The Wit.ai product lets developers add a few lines of its code to instantly build in speech recognition and voice control. Without it, developers would need the expertise, time and resources to build a whole voice-recognition system themselves.
“Facebook has the resources and talent to help us take the next step,” Wit.ai wrote in a blog post about the acquisition. “Facebook’s mission is to connect everyone and build amazing experiences for the over 1.3 billion people on the platform – technology that understands natural language is a big part of that, and we think we can help.”
When a small group of tech people started to praise Amazon Echo, you don't call it a trend yet, but you certain listen to what they have to say, according to
M.G. Siegler,
Admittedly, I’m less than a week into using the device. And honestly, I’m still not 100 percent sure how we’ll use the thing day-to-day.¹ But I now believe in the power of Alexa, Echo’s female voice, as a platform.
In fact, I think Echo makes it very clear that Apple (and to a lesser extent, Google) dropped a ball here. This is exactly how Siri should exist in your home. And this is what that orb thing Google made a few years back should have been.
Connecting users with AI on a personal level, or risk creeping them out, this is the question, according to
Peter Sweeney:
Personification is a powerful creative device. When we attribute a human form or personality to things that are not human, it can make the product more accessible and easier to understand.
In the realm of intelligent assistants, however, this strategy is risky. Like lipstick on a pig, the risks surface not only in how these machines are represented, but also in how they behave.
