What happens when you artfully blend storytelling with artificial intelligence? Done right, you might get a conversational bot smart enough to amuse and captivate us humans.
Take it from animator, technologist, and former Pixar CTO Oren Jacob. Personally picked by Steve Jobs and Ed Catmull to lead the award-winning studio’s internal tech team, Jacob helped craft the moving narratives of Toy Story, Toy Story 2, A Bug’s Life, and Finding Nemo. Now he spends his days imbuing computers, devices, and toys with the gift of gab.
In 2011, Jacob cofounded PullString, a leading player in the “computer conversation” market which offers a bot development platform for text, voice, and video interactions. PullString’s name is partly inspired by traditional talking toys like Woody, one of Pixar’s most beloved characters, and partly inspired by the company’s product philosophy: bots should not be ill-defined, open-ended experiences. Instead, their behavior should be predictable, like the movements of a pullstring toy, designed to produce maximum value for a specific use case.
PullString technology powers many popular products, such as Mattel’s Hello Barbie doll and dollhouse, Activision’s Call of Duty bots on Facebook Messenger, and the Amazon Echo Alexa Skill for popular TV show “The Grand Tour.” I recently sat down with Jacob to learn his secrets behind successful bot experiences and his predictions for the exploding conversational AI industry.
Amazon Echo Killed Costco & iTunes
Jacob has nostalgic childhood memories of following his mother to Costco and playing on the huge pallet pushers in the store. He continued this tradition with his own family, up until the day they got their first Amazon Echo. “Costco vanished out of my family within the first month of unboxing an Echo. We haven’t driven there in over a year,” he reflects. The convenience of asking Alexa for anything and having it show up on his doorstep killed the need for mundane store runs.
His kids have figured out the trick too. “As soon as you wire up your Prime Membership, they’ll try to order ten pounds of something. Thousands of dollars of commerce have shifted from Costco to Amazon. The Echo meaningfully moved the purchasing habits of my family of five.”
Jacob grew up with vinyl records and CDs, while his children have only known iTunes. “ITunes and the digital music market overtook the CD market progressively,” he recalls, “We’ve been a digital music family for over a decade, but we stopped using the iTunes apps as soon as we got the Echo. Music requests and jockeying went straight to voice and never went back. I’m surprised how sudden that was.”
Secrets Of Compelling Computer Conversation
“Whether or not an AI system can pass the Turing Test is irrelevant,” affirms Jacob, debunking the common mythology that a chatbot needs to be humanlike to succeed. Instead he encourages developers to design for “human fidelity conversations,” where a bot is able to engage a human about a specific topic of mutual interest, rather than attempt to hold an open-ended discussion about anything.
“Children have four-hour long conversations with Hello Barbie,” Jacob cites as an example. The doll only has 8,000 lines of dialogue and is a scripted experience. As a crotchety grownup, my own interactions with Barbie were disappointing. Whenever I brought up the subject of AI research or gender inequality in STEM careers, she invariably redirected the conversation back to fashion and shopping. But to Jacob’s point, the doll wasn’t exactly designed for me.
He clarifies further: “the question is not whether I can trick you into thinking a bot is human. The question is whether I can engage you on a particular topic in a way that is comparably compelling as speaking to a human who’s also interested in the same subject.”
In Jacob’s view, the best use cases for bots are scenarios which involve step-by-step agendas, such as following a recipe or building a LEGO castle. Another key advantage of using an Echo or comparable device is that “computer conversations are adjacent,” he says, referring to the fact that mobile phones distract and consume our visual attention, but ambient devices like the Echo can be operated and experienced even in the midst of active group conversations.
“We don’t allow phones at the dinner table,” he emphasizes, “but asking voice assistants about current events is how my kids get their news. They didn’t attach to cable news or newspapers like I did, but they did attach to audio playback. The voice experience was a natural integration into the dinner conversation for us.”
The Battle Of The Bot Platforms
Google, Amazon, Microsoft, IBM, and Apple are all vying for market share in the conversational AI market, expected to balloon to a multi-billion dollar industry. The competition is a boon for PullString, which integrates with all major platforms. “Many successful efforts in the field makes it a rich place for us to do business,” Jacob exclaims.
When asked about their comparative advantages and disadvantages, Jacob draws analogies to the mobile app wars. “Apple created a tightly controlled platform, while Google made Android more open. Each of the major conversational platforms – Siri, Cortana, Google Assistant, and Alexa – reflect their larger company philosophies when it comes to delegation and control between the operating system and third-party developer experiences.” These choices drive whether developer ecosystems, widely believed to be key to platform dominance, will thrive or die.
It’s too early to pick a winner, but PullString works closely with each platform and notices their unique progress. “Microsoft goes above and beyond in developer support. They are attentive and responsive to platform feedback and improvement. Amazon’s conversational offerings have improved the fastest in terms of precision recently. Google naturally dominates voice search as they dominated web search, and I’m impressed at how far they’ve brought Home and Assistant in such a short time.” When asked about the tight-lipped company in Cupertino, Jacob admits, “I honestly don’t know their future plans.”
Jacob is confident the computer conversation market will one day exceed the mobile app market. People often forget the iPhone did not ship with an app store, but breakout mobile successes like Angry Birds and Instagram quickly changed paradigms and inspired explosive growth. He believes the same pattern will occur with bots.
Use of language is also fundamental to being human. “We ask more questions of each other than we ask of Google,” he points out.
In the near future, we might all be speaking to algorithms.