A conference panelist is asked a question. If they answer correctly, they increase their credibility, answer profoundly and they begin establishing themselves as a luminary. Deferring to a co-panelist is basically zero-sum, but a wrong or noncommittal meandering answer could destroy any credibility established during the prepared remarks.
At a cocktail party, a guest gambles on a non-sequitur. If well placed, it can spur interesting novelty in a stale conversation. If mistimed, it will derail the flow and make the guest look disengaged, rude, or self-centered.
Throughout the technological world, researchers and entrepreneurs are working to understand situations like these, to codify and to quantify them. Quantification efforts seek to measure, for example, a panelist’s “credibility score” or a guest’s “conviviality score”. Put another way, to re-imagine such situations as a sort of game, where participants are seeking to maximize their score on one or more dimensions.
Today, learning to navigate conference panels or cocktail parties is only possible through first-hand experience. Each social interaction is unique, and it takes years of attendance to intuit the patterns that enable one to correctly time the humorous aside. However, by codifying situations researchers can aggregate the experiences of many people, and by scoring outcomes quantitatively, to compare strategies. Score-keeping allows for competition, and not just the “leaderboard” or “superbowl” kind, but more significantly, the “Darwinian” and “market-forces” kind.
Consider app-based fitness trainers and the seemingly-simple “3-week streak”. Nike Run Club, Fitbit and the like reward users with badges when they exercise 3 times in a week, 3 weeks in a row. These numbers are no accident. By compiling the experience of millions of users, app providers have determined the inflection point in usage. Users that follow the 3×3 pattern are much more likely to exercise in the fourth week. Encouraging users to exercise 3 times a week is a strategy for building a long-term habit of exercise.
Such strategies, evaluated over a longer time frame, can be controlled for factors like gender, age, time, and weather, combined with additional reward and penalty systems and compared to one another. New strategies can be tested, with novel reward and penalty systems, across large populations of exercisers and aspirational exercises. Winning strategies emerge, “naturally selected” from the infinite array of possible motivational techniques.
In some cases, there are overwhelmingly successful strategies such as the 3×3 exercise habit builder. These become so ubiquitous as to seem trivial. But large datasets also allow for more personalized or situational strategies. For example, how best to engage a regular user and avoid monotony and boredom? Time-challenges and strength building programs in Nike Run Club, for example, draw on each individual’s running speed and duration, age and weight, and self-described goals to present a unique set of challenges, each tied to rewards and penalties in the App.
How then, to handle the tough question at the conference podium? Natural language processing and semantic expert systems are not yet capable of codifying these contexts; computers literally can’t understand what is being asked in real-time, and genuine knowledge of a topic remains beyond the reach of software. But we are beginning to quantify the outcomes, for example measuring the Twitter and LinkedIn activity or the website traffic of a conference attendee. And computers can analyze the natural language in news articles and blog posts that happen after the fact. Multiple companies now provide competing “sentiment analysis” tools that combine these data sets to compile a “sentiment score” reflecting the positive/negative impressions and trending of brands, people, or public policies.
Record labels have long used software to analyze musical recordings. Demo tapes are parsed to identify potential popstars, and studio recordings are regularly remastered based on software’s recommendations to maximize sales. After all, “revenue” is the score that matters most to business.
Gamification is the codification of human situations and strategies of human feedback, combined with the quantification of outcomes, enabling comparison and competition. Individuals identify winning strategies more rapidly with gamification by sharing experiences. Innovation is accelerated by testing novel strategies across experiences through deployment to large populations or through simulation against pre-recorded contexts. In short, gamification enables both humans and computers to learn, and accelerates learning.
Big-data, machine-learning and gamification are mutually supportive technologies, all driving the fundamental benefit of accelerating learning. No mere tech fad, we should expect these three trends to drive wide-ranging and potentially transformative changes. Personally, I can’t wait to attend an AI conference with a nonhuman panelist. In the meantime, share your human thoughts with us here.