So You Missed the Humanity.ai Conference—Is Your FOMO Founded?
I regret to inform you that yes, you missed out. I’m sure your “work obligations” and “responsibilities” were well-deserving of your attention, but the best we can do to assuage your latent social anxiety is offer you a recap. Strictly speaking, Humanity.ai was a single-track (no workshops) conference created to foster an industry-wide conversation on human nature in the midst of artificial intelligence, and artificial intelligence in the midst of human nature.
So what did that conversation sound like? It was as if your imagination—encapsulated in a clean, clear forensic examination box (adorned with a corporate conference lanyard badge, obvi)—rested innocently in a folding chair while theoretical neuroscientists, Pixar creatives, Machine Learning experts, and science fiction scholars took turns poking at it with sharp-yet-tickly sticks made of FUTURE REALITIES. So frkgn rad.
Not sure if that’s helping your FOMO. #sorrynotsorry
Well if you had attended, you probably would have arrived seven minutes late — uncannily, just like me — and found a seat just as Capital One’s Steph Hay (Head of Content, Culture, and AI Design) and Brandon Schauer (Head of Emerging Design Practices) took the stage to usher us into the brain-space of Here We Are vs. What If? Their nervous thrill was palpable, but hey — their dream robo-design-meta-science-non-fiction conference was about to actually happen after nine months of planning, and there’s a lot of room in my heart for that kind of jitter.
The audience of 200 seemed to be made up of designers and engineers of varying breeds, unsurprisingly. What did surprise was the unique angle of each presentation. No repetition, no rehashing of discipline standards, and nobody left early. The interwoven topics of AI and the human experience as a thought project is still new to the landscape of corporate conferences, and the pairing allows for an incredibly broad range of perspectives. There was too much to cover in a single day, so here’s hoping Humanity.ai 2018 offers an expansion pack.
You can check out the @Humanity_AI Twitter feed for the play-by-play, and/or read below for detailed reflections — ripe for stealing when you pretend to your friends that yes, you did get to attend that amazing conference.
How to Robot-Proof Your Kids
Vivienne Ming (Founder & Executive Chair — Socos)
The depths to which Vivienne sees into the future is simultaneously staggering and motivating. She flips through third and fourth-degree implications of her curious work in AI with the ease of thumbing through a magazine in which all of the headlines make you O.O a little: Facial Recognition Used to… Improve Refugee Camp Experience? Create a Bot to Automatically Send a Text When…. Your Child’s Blood Sugar Level Drops? Her ultimate thesis: We must transform education to enable craftsmen of the future.
That craftsman, according to Vivienne, will be a creative, adaptive problem-solver…which is more than a tagline for your LinkedIn profile. The education we’re offering kids now won’t serve them by the time they grow up when even middle-class, college-requisite jobs have been outsourced to automation. Robot-proofing our kids means making them more uniquely human.
Constructing a World of Conversations from Scratch
Oren Jacob (Co-Founder and CEO — PullString)
“Character is not algorithmically derived.”
“Conversation = content AND intent.”
“Conversation is the essential human creation.”
Oren’s presentation was effortlessly engaging, but he spoke so quickly it was tricky to absorb and record more than bullet-points. Aaron Sorkin references in tow, it was the sort of blazingly fast and funny presentation that you just had to be there for.
Trends in AI: Where We’ve Been, Where We’re Going
Journalist Panel: Ellen Huet (Bloomberg), Khari Johnson (Venture Beat), John Mannes (TechCrunch)
Audience question: “As we build more and more autonomy into our technology, if/when AI is able to make decisions that put humanity in danger, who should hold the master key? Governments, corporations, the United Nations?” The question we’d all been waiting for.
John: “We need to drop the concept of AI destroying the economy, the take-over mantra.” Here we learn that John is an avid Skynet Denier. “It isn’t beneficial to ask this question… it isn’t going to happen.”
Khari: “The bigger concern is humans, now.” He doesn’t say it won’t happen, but suggests we use our creative energy to solve more immediate problems.
Ellen: Referencing zombie contamination a là The Walking Dead, she offers enigmatically, “It’s a very dangerous assumption that if it’s been touched by a computer, it’s inherently fair or unbiased.”
Chatterbots: Content Strategy for the Conversational Interface
Elena Ontiveros (Content Strategist — Facebook)
“The faster the people in this room succeed, the faster I’m out of a job.” That wasn’t Elena speaking, that was my internal monologue repeating ad nauseam as I watched her detail the word-for-word creation of the chatbot used at Facebook’s F8 conference.
Full disclosure: I’m an executive assistant by day / content strategist and storyteller by side-project, and Elena’s work was perhaps the most motivating presentation of the entire conference. More of a stick than a carrot, but I’m no dum-dum. Alexa, Siri, Eno, Vladimir — one of them, eventually, will take my current job.
AI won’t replace executive assistants until the robot can tell you “No,” while making you think you heard “Yes,” — but Elena’s extraordinary work is a prime indicator that that day isn’t far off.
Removing the Sociopath from AI
Mark Walsh (Animator — Pixar, Motional.ai)
My inner monologue continued chanting while Mark gave us a tour of the Uncanny Valley and described different types of servants we might eventually mold the robots into. It’s ironic to find yourself rooting for the technological advancements that will eventually lay claim to your paycheck. Something whimsical, yet stilling… Gallows humor? Is this what my Subaru-mechanic-husband feels like when he works on a Tesla?
Well Mark, for his part, was fantastic. With all of the vibrancy one would expect in a Pixar animator, he named personality, empathy, growth, and trust as components of humanity that must be developed within AI in order overcome the currently-inherent sociopath. My inner turmoil notwithstanding, would watch again.
Lessons Learned from Birthing a Bot
Chris Messina (Developer — MessinaBot)
In an account of what it’s like to create a customized, personal chatbot which will communicate and make appointments on your behalf (“The faster the people in this room succeed, the faster I’m out of a job,” ahem), Chris named awareness as one of the greatest challenges facing bot development. Not the bot’s self-awareness, but humanity’s awareness OF bots — they’re still fairly unknown to the consumer at large. How to make them, how to use them… the frontier remains predominantly unmapped.
Chris focused on functionality and specific mistakes (i.e., discoveries) he’d made, and I appreciated his candid curiosity. He seemed genuinely happy to keep poking at his own ideas to see how he might iterate in the next version.
Using Bots to Grow Relationships at Work
Veronica Belmont (Product Manager — Growbot), Jeremy Vandehey
(CEO — Growbot), Amir Shevat (Director of Developer Relations — Slack)
This panel on Growbot—a Slackbot that celebrates team wins and encourages appreciation between peers at work — laid bare the power of Content Strategy.
The casual, rapid-fire nature of Slack conversations lends itself to humor, memes and inside jokes — but what happens when your chatbot’s one-liners fall flat, or worse, come off as patronizing? It’s hard to generate genuine rapport with an entity that seems to be rolling its eyes at you.
What was Growbot’s magical solution to inadvertently offending its users? Test your content. Revise and test again. Develop your content strategy practice (beyond taking a workshop — bring in the experts) to master the language, story and voice of your products.
Odds and ends from the notebook:
“Don’t fake intelligence.” Okay, check.
“Keep bots domain-specific for best success.” Excellent news for my job security because executive assistants need to know about EVERYTHING.
“Read The Circle.” Just told Alexa to add this to my book list. … The irony does not escape me.
Bots for Social Good
Josh Browder (Founder — DoNotPay)
Josh has helped many thousands of people appeal traffic tickets, manage their prescriptions, get reimbursed for flight delays and fight unfair evictions — for free. He coined the quote of the day (“If chatbots are going to help humanity, they have to do a lot more than order a pizza,”) and as if his inspiring work wasn’t enough to make you root for him, Josh has the most unassuming, natural presence on stage. Entirely absent was any over-practiced flair of a TED talk. He comes across as demure, reserved, but assured, clever, focused, genuine. Would easily watch again.
AI in Film & Storytelling
Moderator Chen Shen — Capital One), Jay Shuster (Production Designer — Pixar), Karen Dufilho-Rosen (Executive Producer — Google Spotlight), Rebecca Stockley (Co-Founder, BATS Improv), + Mark Walsh (CEO—Motional)
My notebook reads, “Somehow we jumped right into Datalore and Westworld.”
These five AI-loving creatives bantered about their favorite fictional characters, assigning gender to virtual assistants, anthropomorphizing technology, the charm/terror threshold, the delineation between audiences who want to participate in vs. observe a story, and even included a first-hand account of the Oscars snafu. There were guided questions and answers, but no agenda, which made this conversation all the more natural and all the less replicable.
Recognizing the Cultural Bias in AI
Camille Eddy (Machine Learning Practitioner — HP and Boise State)
Transparency is key, according to Camille. Throughout the presentation that referenced several thought leaders in the Anti-Bias digital space, Camille introduced us to Explainable AI (XAI) which advocates for clarity between the user and program. Users — not just developers — need to understand why an algorithm gives us a specific response to a given set of data. I walked away wondering if clarity would be most widely achieved by a) dumbing down the explanation, b) elevating the user’s knowledge of data science, or c) creating an intermediary to translate between the two? Is XAI in fact a new language both the bot and lay-user could learn in order to increase transparency?
Critical Tour of AI in Sci-Fi [Film]
Chris Noessel (Global Design Practice Lead — IBM)
Academic sci-fi fans, this was the talk for you. And me. Chris detailed seven AI tropes in science fiction films with impressive depth (Oracles, Domestiques, Apprentices, Assistants, Cops, Bulldozers, Murderbots), and then gave us the skinny on what’s being whispered in Algonquin-style robo-ethicist circles these days:
Bulldozers (AI which dismisses human safety in pursuit of a goal, i.e. Mother from Alien) are a more realistic threat to humanity than Murderbots (AI whose primary goal is to eradicate humans, i.e. the Terminator). I wonder, what would Skynet Denier John Mannes say? Hmmmm.
tl;dr — Heady questions ruled the day.
Are we focusing on problems in the technology, or on the human problems technology is supposed to solve?
Are we cloning ourselves, or augmenting our existence with superpowers?
In the wedding story of Technology and Humankind, is artificial intelligence a protagonist, antagonist, MacGuffin, or prop?
And is it even possible for these answers to be prescriptive, or are they necessarily descriptive?
If you’re a FOMO reader, don’t wait until next year to join the Humanity.ai conversation. Google these speakers, read their wares, and send us your thoughts!
If you did attend the conference, what were your favorite parts? Whose books and articles have you already devoured and pressed on friends? Have you Facebook-stalked any of the presenters? Respond in the comments or send me an email, I’d love to hear.
Photos by Amy Tolbert
