For decades, commercial fleets have been using telematics to track their vehicles and in more recent years to measure and report unsafe driving behaviors like harsh acceleration, braking and cornering. Most would agree that fleet telematics was the first wave of fleet safety technology.
The next wave is currently under way. Many commercial fleets are starting to deploy dash cameras (dash cams) as a complementary solution to telematics. When they were first introduced, dash cams were pretty basic but now they have evolved to using AI to identify, warn and report risky behavior. The “seeing capability” powered by AI provides edge computing capabilities, which can interpret visual images and determine risk exposure for many additional behaviors.
In addition to dash cams becoming better identifiers of risky behavior due to AI, their adoption has also been accelerated by the auto insurance industry due to the fact that the video is extremely beneficial in the area of claims settlement and accident exoneration. In fact, commercial auto insurers have had a huge hand in mandating and/or motivating dash cam adoption by fleets in verticals that were traditionally not receptive to safety technologies due to drivers feeling that “big brother” is watching. What fleets and insurers have discovered is that forward-facing dash cams are still valuable for exoneration and do not carry the “big brother” stigma. For this reason, we are seeing a strong trend of dash cam adoption that is FORWARD-facing vs. DUAL-facing.
While AI and auto insurance have combined to accelerate this second wave of fleet technology adoption, companies should be cognizant that a camera is only as good as its field of vision. Just because risky activity can’t be seen due to dash cam blind spots doesn’t mean the risky activity didn’t happen. While some risky behavior is hard to hide like drowsy driving or eating, other activities are easier to obfuscate like sneaking a phone call or a text message.
Easier activities to identify. Some of the common capabilities offered now via AI dash cams include being able to see the driver, where the driver’s attention is, if a seatbelt is being worn, if the driver is using a cell phone, if the driver is following the vehicle ahead too closely and if the driver is running stop signs. All of this, in addition to the previously available insights regarding aggressive acceleration, braking, cornering and speeding, combine to provide a safer driving environment.
Harder activities to identify. The issue is that some of these AI-derived elements are more challenging to determine and the accurate identification of such risky behaviors is compromised. While some dash cam providers claim to see a 90% reduction in cell phone use by providing an immediate audio warning to the driver about identified phone handling, the challenge is really in the identification of the behavior. The fact is even if 90% of identified phone events can be corrected, this still begs the question as to what percentage of phone events are outside the camera’s field of vision and NOT being identified? Dash cam blind spots exist because the camera simply cannot see everything. Employees using their phones behind the wheel can be difficult to catch for some of the reasons below:
- The cell phone is challenging to identify because it’s not attached to the driver’s body and the driver can simply move it to an area outside the camera’s field of vision.
- A driver can place a phone below the mounted dash cam or can hide it in the lap.
- Phones held in the left hand tend to be more difficult to identify than phones held in the right hand (nearer the direct line of sight of the dashcam).
- If the driver is holding the phone in front of his body, it is challenging for AI to determine if what is being held is a phone or the driver is eating a snack, say a yogurt container or drinking a can of soda.
- These situations can easily lead to false positives, declaring a phone handling event when none was present.
- These false positives annoy drivers and managers who have to pay attention to them.
- The AI dash cam identification of cell phone use (which requires analyzing the driver) may also present the fleet with issues of privacy, not only driver concerns, but also union or management concerns, or concerns about the legality of use of facial geometry, especially related to driver identification. These privacy concerns have led companies to create clothing for privacy fans that have been designed with patterns to confuse AI and prevent people from being identified by AI-powered cameras.
Legal Resistance. In addition to whether cameras are comprehensive in identifying particular driver behaviors, there may be a larger concern looming down the road. Drivers as well as unions often object to the use of video technology which they consider to be an invasion of privacy. Certain states have passed privacy legislation that prohibits, among other things, the collection and dissemination of biometric data without consent. This type of legislation would also prohibit the use of AI-generated data for identifying driver distraction without obtaining the driver’s consent. Recently, a major dash cam provider settled a class action lawsuit for the alleged violation of the Illinois Biometric Information Privacy Act where machine vision and artificial intelligence was used to predict distracted driving events. While AI applications appear acceptable for all road-facing situations, the issue of recording the driver is a clear concern. The potential adoption of similar privacy legislation across the U.S. could lead fleets to choose road-facing cameras in the future.
The Coaching Hurdle and Dash Cam Blind Spots
One of the most positive trends with AI dash cams is the identification of a risky behavior accompanied by a near-immediate audio warning to the driver. The driver may have a period of time to self correct and if that happens, the behavior may not be recorded as an “event”. This concept of “self correction” is a key to moving fleets in a safer direction while actually lessening the need for extensive driver coaching and management review.
Cell phone use is an interesting case since it is the one behavior that truly is more of an addiction. We are drawn to our phone whenever it calls us or beeps or buzzes. We may still see drivers using their phones despite audio warnings from the dashcam.
Like other risky behaviors, those that continue to show dangerous behaviors may require a coaching session to emphasize the importance of the issue and the path to correction. Fleets vary greatly in their commitment and execution of coaching, often affected by geographic focus or the presence of a particularly good safety manager. Driver behavior coaching gets stale over time.
With the importance of cell phone distraction as a key contributor to fleet vehicle accidents and the unique, addictive nature of cell phone use, it is critical for fleets to do whatever they can to mitigate the risk of such behavior.
Combining both a dash cam with a cell phone compliance solution provides the best defense to protecting your company from collisions due to distracted driving.
- If your fleet has a union, is getting resistance from employees due to privacy or wants to avoid the smoking gun video that wasn’t effectively coached, you should still consider forward-facing dash cams to provide the exoneration benefit + cell phone compliance for claims reduction.
- Forward-facing dash cams + LifeSaver Mobile is a perfect combo.
- If you can overcome union/employee resistance and believe that you can effectively coach most of the video events captured by your AI dash cam, you should consider dual-facing dash cams + cell phone compliance for claims reduction.
- Dual-facing dash cams + LifeSaver Mobile is also a perfect combo.
