This week, AAA released their latest round of distraction research which focused on hands-free in-vehicle infotainment systems.  In one study, 10 different integrated systems were evaluated, each over the course of 1 week by a new user.  In another study, 3 different voice-activated assistants (Siri, Google Now, and Cortana) were evaluated in a more controlled session.  Results compared cognitive distraction levels of a voice-controlled task to a series of researcher-defined passive tasks, and a very mentally demanding surrogate task.  Based on this very broad comparison, the study authors conclude that all hands-free interactions are mentally distracting for a driver, and in some cases have residual distracting effects even after a task has been completed.  Differences between systems are discussed as well.

AAA has sponsored and published many similar research efforts over the past few years.  Each has roughly the same methodology, which compares the mental workloads of various in-vehicle tasks to a broadly-defined baseline and an unrealistically distracting surrogate task.  Each of these efforts has led to roughly the same conclusion: In-vehicle tasks (performed manually or hands-free) can tax mental workload while driving.

However, despite many years and several iterations on this line of research, AAA and related advocacy groups have yet to establish a true statistical correlation between mental workload for specific tasks and a direct measure of safety.  Meanwhile, rigorous naturalistic research from the Virginia Tech Transportation Institute (Dingus et al, 2011) shows no increased crash risk for cognitive tasks like phone conversations.  Additional real-world research published in Risk Analysis (Young & Schreiner, 2009) shows that hands-free conversations do not lead to a significantly increased crash risk.

Without a specific and demonstrable relationship between mental workload and crash risk (or at the bare minimum, an acceptable safety metric such as eye glance time or lane centering), the latest results from AAA continue to shed no light on the impact of mental workload on real-world driving.  As such, these results need to be taken with an enormous grain of salt.

More important to note, though, is that the future best practice for in-vehicle infotainment systems is multi-modal.  As we discussed in our review of HMI trends and best practices, the most suitable HMI for each top in-vehicle task is strongly dependent on the task itself.  A singular focus on creating guidelines for one HMI modality like voice control is short-sighted and counterproductive.

Unfortunately, despite these concerns, AAA’s research activities are dominating the public debate on the effects of in-vehicle systems on driver distraction.  As discussed in our report on a previous AAA research effort, the AHEAD consortium groups led by MIT’s AgeLab are researching the distraction presented by multi-modal aspects of various infotainment systems.  This line of research will prove far more valuable in building a higher-fidelity picture of driver distraction.  However, in order for efforts like this to move forward in a meaningful way, OEMs absolutely must get involved to help drive the public discussion.