Startups Monitor Emotions to Optimize Consumer Value, Vehicle Safety, and Employee Performance

890

Many of the advances in internet of things (IoT)-based data gathering and analytics sound like Orwellian spying. And indeed, nearly every week brings new instances of devices streaming sensor measurements to a cloud-based server (Bluetooth-based smart locks, for example), where vulnerabilities compromise the endpoint, transit, or storage – and intentional system features exploit users’ unintentional or even willful release of data for ethically-dubious ends. Now the addition of affective computing, a branch of artificial intelligence (AI) geared towards sensing and emulating human emotion, is adding a new layer of controversy in spaces like:

  • Enjoying entertainment in smart homes. Affective computing startups are entering the home with social robots like Jibo, but also in ways that are less obvious (or voluntary) for consumers: last year, Samsung had to walk back smart television “spying” features that sent not only voice commands, but incidental conversation back to its servers. Today, Affectiva is working with consumer goods and media companies like DisneyCBSCoca-Cola, and Unilever to measure the effectiveness of ads in sparking emotional responses. The company, which raised $14 million earlier this year, has now launched a mobile-device based version of its software.
  • In traffic, via connected vehicles. GM’s “Valet Mode” car monitoring system provided driver behavior data and in-car video and audio recording to Corvette owners, but also to third-party marketers, and may have inadvertently made its users into wiretap felons. Now, Nexar, an Israeli/San Francisco startup, uses the camera on drivers’ dashboard-mounted smartphones to record and analyze video footage of the scene in front of the car. Unlike dumb dashcams, the app uses AI (machine learning) to spot other drivers’ bad habits – hard braking or near misses, for example – and warns the network of Nexar users whenever a bad driver is nearby. Moreover, the phone’s screen-side camera analyzes in-car events, promising to help make drivers and passengers safer in ride-sharing and taxi services (it claims 600 users in Uber, Lyft, and Via fleets). The company launched the app in February 2016, and raised a $10.5 million Series A in June. Another such app is Nauto, which targets vehicle fleets and insurers, promising to reduce risks by detecting aggressive drivers externally, or driver drowsiness in the vehicle. It raised a $12 million Series A in April.
  • On the job, in intelligent offices. A cottage industry is arising around the use of IoT data to monitor employees’ productivity and social interaction on the job, with companies like Enlighted and Yanzi Networks putting sensors in conference rooms and personal desks to create the “quantified employee” or the “human cloud at work” . Cornerstone OnDemand is a cloud-based big data and analytics company focused on office productivity, and ThingLogix has extended its business operations software to include employee-monitoring devices. As Deloitte wrote after an experiment putting sociometric badges on employees, “With oceans of data from workers’ wearables, HR departments could aim to create more pleasant and efficient work environments by looking at productivity, patterns of communication, travel and location trends, and how teams work together.” While much of this tracking relies on objective measurement of tasks and interactions, at least two new startups are adding measurements of emotion to the mix. Humanyze offers “people analytics” based on data captured by wearable “GEM” badges (the same ones used by Deliotte) that use microphones to monitor the “tone of your voice and how frequently you are contributing in meetings,” while “hidden accelerometers measure your body language and track how often you push away from your desk.” Another company, Behavox, applies machine-learning algorithms to voice and text communications, as well as employee task data (like traders’ transaction records) to spot compliance problems in financial services, and “highlight both good and bad conduct.”

As with many new technologies that tinker with cultural norms and assumptions, the use and abuse of affective computing (like social media, online dating, and even email before it) will seem creepy to some, but efficient and practical to others (see the Lux Research report “The Future We Want: Directing Customer Behaviors with Predictive Analytics and Nudge Technologies“). It seems inevitable, for example, that workers should wear intelligent monitoring gear that can help protect them and others in areas where safety risks are high. Over time, the practical camp is likely to outgrow the resisters in emotional areas as well, while the measurable and monetizable benefits grow and grow. Readers are advised to be fast followers – not leaders or laggards – in the affective space, developing or applying these technologies so they can reap the benefits of better understanding of consumers and employees, while not becoming the poster child for invasive practices.

by Mark Bünger, VP Research, Lux Research