An anthropologist’s take on conversational interfaces
One of the biggest current topics in technology is the rise of the conversational interface, and a huge part of that discussion is the impact that these interfaces will have on the economy — from individual jobs, to ecosystems, and venture capital. As an anthropologist, I’m interested in the potential cultural implications heralded by the undergoing shift in technology.
Since language was invented, humans have undergone shifts in communication with one another. For a long time, most communication was strictly face-to-face; then letters became a way of corresponding with those far away and arranging meetings with those nearby. With the invention of the telephone, speech dominated writing once again. The rise of the personal computer and widespread internet access led to an increase in text-based communication as e-mail became a primary form of workplace interaction, both in and out of the office. With the dominance of mobile devices and messaging apps, our communication has become primarily text-based, perhaps for the first time.
Increasingly, when we are at work or in public, we are simultaneously inhabiting two worlds: our physical environment and the world of text. The physical world requires immediate responses, and is where most of our verbal conversations occur; if someone is in the same room, it’s usually easier to just flag them down and ask them something than to text them. But for anyone who isn’t immediately available in the physical world, texts let us communicate our thoughts as they occur to us, quickly and easily. Although various norms exist around response times, text doesn’t necessarily demand immediate responses the way the physical world does. Text-based communication enables us to do research, check with others, and prepare before responding. In giving us that time to respond, it also allows us to maintain relationships and communicate regularly regardless of the time zone they’re in or the distance between us. Texting also lets us more easily schedule and plan to get everyone on the same page — and in the same physical space. However, even with the obvious benefits of conversational interfaces, the immediacy of texting means that we’re faced with a firehose of communications, inundated with messages all day long.
A similar shift is occurring with our use of intelligent assistants and the internet. We are often more comfortable typing into our laptops or phones than talking aloud to voice assistants, at least in the public sphere. Voice-based assistants like Siri or Alexa can be very useful when in the car or when our hands are occupied with other tasks (washing dishes or folding laundry, for example), but in office settings and in public, text-based assistants are a more practical choice.
As we increasingly engage with the text-based, digital world, the software and tools we use are beginning to learn our preferences and streamline our interactions with both this virtual world and the physical one — if I search for furniture, furniture companies are able to target me with specific ads on Facebook. I research the Elizabethan era for an extensive project, and no matter how many Tom Watsons there may be in Wikipedia, Google knows to put the one born in the 1550s at the top of my results page; it starts to know what cities I frequent and what kind of books I like; Netflix narrows my preferences all the way down to “cerebral supernatural independent dramas with a strong female lead”. Facebook still vacillates between thinking I’m engaged (wrong), looking for love on a Christian dating site (doubly wrong), and looking to adopt a child (nope), so we’re not quite in an era of perfect tracking of every individual’s preferences — yet. Intelligent assistants, too, will take cues from behavior to cater to a user’s needs.
Conversational interfaces are set to combine these two trends. Digital assistants which primarily use text, “living” inside the texting and messaging apps we already use, and which learn our habits and preferences as they go, bring with them the possibility of major advantages, especially in the workplace. Eliminating the busywork of logistics by automating it will free up more time for higher-level work, and for further creativity and innovation — or for binge-watching Netflix. The bots, which work within workplace messaging systems such as Slack, waiting to be given a task, or proactively reaching out, can do things like schedule meetings, book conference rooms or make reservations, and then send a message with the time and place to everyone involved. By knowing all everyone’s schedule and preferences, our bots can be decisive for us. This not only saves time, it saves frustration; if everyone’s busy, dealing with a particularly difficult project or having a long day, that little bit of streamlined scheduling can go a long way. Even under the best circumstances that extra efficiency helps everyone get things done faster and maintain focus on the project at hand. So far, scheduling and workflow management seem to be the two places where bots are emerging in the workplace, but as the technology develops, we’ll continue to see more and more powerful assistants become available, continuing to change the way we work and interact.
As conversational interfaces continue to learn and act as intermediaries between users and the rest of the internet, they may continue to tailor each individual’s experience of the digital world to fit their preferences. These integrations, whether with search or through other tools, can increase efficiency and output by enabling workers to spend the majority of their time within a single interface, rather than entering and exiting multiple platforms throughout the day. However, a seamlessly individualized experience of the internet is not without possible consequences: for example, a loss of serendipity, as we already see in Netflix — it’s much harder to stumble upon something weird and new than it was wandering through a Blockbuster (R.I.P.), though Netflix is also obviously much more convenient, and most people would agree that that loss of randomness is a price they’re willing to pay. In decreasing the exploratory nature of the internet, we might see another unintended consequence: a filtering of knowledge that could ultimately result in people only finding evidence that supports their predetermined views, and encountering fewer differing opinions and experiences. Because so much of the early promise of the internet was in the possibility of interacting with people all around the world, with very different views and backgrounds, we must be careful to ensure that we don’t allow that possibility to be filtered out. The impact bots might have on that still is unclear.
On the other hand, for every person who used the internet to find their tribe or expand their worldview, there is someone who used it to cyberbully or express hatred, and find confirmation of their narrow view. Like any other innovation — and like the internet itself — what happens with conversational interfaces will depend on how we use them.
