They say the Lord moves in mysterious ways. When God wanted to speak to humankind, he didn’t speak directly. He used a conversational user interface (CUI).
And the angel of the Lord appeared unto him in a flame of fire out of the midst of a bush: and he looked, and, behold, the bush burned with fire, and the bush was not consumed.
And Moses said, I will now turn aside, and see this great sight, why the bush is not burnt.
And when the Lord saw that he turned aside to see, God called unto him out of the midst of the bush, and said, “Moses, Moses.” And he said, “Here am I.”
Conversation is the original interface. It worked for Moses. And it worked for Ali Baba, who in The One Thousand and One Nights, used a CUI and the phrase “Open Sesame” to gain entrance to the riches behind a magical cave door. In our literature, our sacred writings, and our folklore we’ve imagined CUIs and bots for a long time. In the 16thCentury the Rabbi of Chelm created a bot out of river clay, and speaking to it, brought it to life so, according to a Polish Kabblaist, “It performed hard work for him, for a long period.” Of course, he didn’t call it a bot. He called it a Golem.
Today, conversational bots are proliferating wildly, including Alexa, Siri, AskGoogle, and in films: 2001’s HAL, Samantha from Her, C3PO, and R2D2, etc. Real, virtual, or imagined, we’ve been speaking to bots for a long time already. These represent the front-ends of AI and the systems that control them.
In the process, we invest bots with agency, personality, and even emotional states of being. Add a voice to a piece of code, engage in a conversation, and suddenly we are extending empathy to software. We watch HAL transform in 2001 from a helpful system then, as his agenda is revealed, as scheming, and as his code unravels and he sings the nursery rhyme he was first taught (“Daisy, Daisy,”) we are touched by sadness. That’s a lot of emotion to extend to a blinking red light. Conversation nearly immediately triggers us to see character in code. Add conversation and we move from prop to personality.
Add a voice to a piece of code, engage in a conversation, and suddenly we are extending empathy to software.
Conversations invest bots with character and emotion. If that’s not godlike, it’s at least person-like (which is still a big step). When the product or service has a human quality, in our eyes they become human or human-ish. Imbue them with new technologies like machine learning and their capabilities become superhuman. Right now you can access all of the knowledge on Wikipedia or buy all of the products on Amazon simply by talking to Alexa. No human can do that. Talk to your home bot and it will turn on the heat in your home from a continent away.
At Seed Vault Ltd. we’re fascinated by opportunities to use CUIs to help us compare insurance programs, diagnose aches and pains, access our medical histories, process purchases, and provide guidance to our financial investment decisions. Just like a Golem, we expect bots to do our bidding, and “perform hard work for a long period.” But, we also see them as something more — whether that is agency or influence or an emotional extension of our own being.
These are complex systems to build well. The good news is that the hard work by a developer can be repurposed, reconfigured, and resold. For example a healthcare bot could be licensed over and over again. Kaiser, Aetna or Blue Shield would have the opportunity to customize the bot for their own organization, use 90% as is, then reconfigure the rest for content and personality. In this process, they accelerate development to weeks instead of years. In addition, it’s possible that developers could use components of a healthcare bot (like the part that probes for information) for other kinds of bots (insurance, roadside assistance, financial services, customer service, or calling for help with a plumbing disaster on Thanksgiving). Bot code and services, then, become the components of a marketplace of valuable, reusable, and much-needed software. Such a market economy that incentivizes developers to create bots, in which bots compete on price, quality, functionality, would accelerate the creation of many, many more bots in our lives.
We’re still on the shallow end of this curve. Being helpful and turning lights on and off is adolescent stuff (if your adolescent is an eagle scout). Bots are advancing past these kinds of features quickly. But, like teenagers, bots don’t yet have great judgment. We need to help them be more socially and functionally successful as well as build a platform for them to be trustable. In time, their capabilities will grow to be much more significant, making the rules of their engagement (and the platform they run on) all the more important.
As bots mature, we’ll see them outgrow their clunky adolescent phase and mature into even more useful adulthood. It won’t be long before we have triage bots than can pass your medical information to the right nurse or doctor. The bot might not create life, but in an emergency, sending the right information to the right person could save a life. Not godlike, but close.
Nathan Shedroff is the CEO of Seed Vault LTD, which is building the Seed Token project. A pioneer in the fields of experience design, interaction design, and information design, he is also the chair of the Design MBA programs in design strategy at California College of the Arts in San Francisco, and author of many books.