AI will change much of our world. There will be opportunities for measurement and evaluation not yet dreamed of. One example has already begun and provides an overview and perspective worth our consideration. There will be more examples as industry and individuals employ tools to their individual benefit.
The news recently provided insight into a device that uses multiple cameras to inspect vehicles. It has an array of cameras and can identify, evaluate, and memorialize conditions. The obvious deployment for this is likely vehicle fleets, such as car rental. But the implications are likely much broader from the perspective of safety, maintenance, and repair.
The news story focuses on the rental car application and provides a photo of a car being driven through an arched scanner of many cameras. The vehicle is thus viewed from all angles and analyzed by an artificial intelligence (AI) tool, and assessed each time it is returned. It is not clear whether the car is scanned as it leaves the facility.
In this instance, the AI tool found a small ding on the car. Perhaps unnoticed or even unnoticeable. The customer was charged for the damage. Was it there when the car was rented? Did it occur during the rental when the vehicle was parked? Did it occur in the rental facility between rentals? Whose responsibility is it?
Suppose the car is scanned only when driven into the rental agency (return). In that case, any damage that occurs to the vehicle while it is in the rental facility might go unnoticed or unnoted until the next time the car is returned. A customer might be more comfortable with the process and the technology if a scan occurred upon departure and return. Or, perhaps the customer must look out for themself? Caveat emptor?
In this instance, the customer was later notified by Hertz of a charge for repair related to a minor cosmetic flaw (ding) that the AI noted. She is contesting the charge and denying responsibility. I have rented a few cars over the years, and it is always possible to incur such charges. Many times, the renter is time-obligated to leave the car and head to a flight, trusting that the vehicle will be checked in and the contract closed.
I almost always do a walk-around video of a car before I leave the lot with it. I never drive off in a moving truck or rental trailer without doing so, because moving can so easily include scrapes, nicks, and dings. Despite my careful engagement, I was charged a pet cleaning fee once, when I never had an animal in the car. I now do a walk-around and interior video when I depart and return.
This illustrates, again, my contention that technology generally and AI specifically may be in an Arms Race (May 2024). The ascent of the AI car scanner shows the potential for more rapid vehicle return and more thorough examination then for potential damage. This may empower the rental agency to both deliver service and avoid inadvertently absorbing the cost of vehicle repair. And to do so without incurring the labor cost associated with a detailed inspection.
Nonetheless, there may be disputes with customers, as illustrated by the news story above. Any customer might take exception to the AI conclusions and fight the rental agency's demand for damage reimbursement. That may be a challenging undertaking in the face of evidence produced by the AI scanner, essentially a full PET scan of the departing vehicle.
Enter the consumer response, reported by the New York Post. There are now phone applications that similarly allow the renter to "create their own ... tamper-proof, AI-powered before-and-after damage scan in seconds.” And, in the realm of evidence authentication, the app "not only identifies scratches and dents but also timestamps, geotags, and securely stores the images to prevent alteration." That is a step above my habit of making videos.
In this, some will see the resolution of a conflict between the car owner and the car renter. Others might see collaboration in that each side merely seeks to accurately allocate responsibility for damage or loss to the person appropriately at fault or responsible.
But, in the broader perspectives of AI and work, the main point that struck me was the loss of jobs. In all of my encounters renting cars, I have dealt with a human who checked me out of the rental facility and another who checked me in. That process is both labor-intensive and time-consuming. A human on either rental or return might merely assure you have taken/returned the correct vehicle, or they might do a detailed inspection. I have not had many rental or return experiences that were detailed.
In such a human interaction, there is room for human frailty and error, perhaps worse. A security guard has been arrested in Miami for allegedly turning the other cheek or facilitating the theft of more than a dozen rental cars, according to WPLG. The rental agency targeted was also apparently Hertz. Humans can fail at an assigned task, and so there is some inclination to technology. Yet, technology can also fail.
My first experience with eliminating some of the human interaction in car rental was an app. About three years ago, a rental company encouraged me to download their app; with it, I can often arrive, identify my vehicle, gain access, and exit the garage all without speaking to anyone. That is admittedly convenient. It has worked swimmingly at some locations and at others, not so much.
Thus, there is no ideal solution to all challenges, but persistence that may require dynamic and periodic assessment and management. Staff may require supervision to do the job (inspect the vehicle studiously) or to prevent malfeasance (stealing cars or helping those who do). Management may have to balance the challenges of technology and the demands of keeping the service or product flowing.
Nonetheless, these AI scanners could be a tool for both the rental and return processes. There is the potential for more rapid completion of both, without the attendant expense of labor. The technology may persistently improve with both the business and consumer tools innovating, upgrading, and competing. This illustrates once again the potential for AI to impact employment.
So, it may be that the scanner technology is not yet universal, or that the reliance on it is less at some vendors than others. Some may make consumption decisions based on their perceptions of this or that vendor's deployment and reliance. If a customer does not like these scanners, they may gravitate to rental agencies that do not use them.
Some vendors may make similar decisions regarding who they will or will not accept as customers. If a customer is engaged in damage disputes, an agency might decide not to rent to them in the future.
Customer and vendor, each with an AI tool, each striving to preserve or project. Each tool strives to accomplish a primary goal but is also persistently revised and improved to answer the demands and shortcomings of the other tools.
The implications are broad and deep. As AI enters the workplace in the role of assessing and evaluating work quality, productivity, and efficiency, might workers similarly respond with their own AI that affords such evaluative assessment at the point of production? In the realm of labor and management relations, it is likely that each will engage and employ such tools in a check-and-balance not dissimilar from the car rental scanner scenario.
Carrying that similarity on, some employers may be less inclined to recruit or retain those who would resist such tools. Some humans may similarly avoid working for companies that deploy such tools. And in any event, it is practical to anticipate that deployment, response, and equilibrium will be dynamic and unpredictable for years as this AI arms race continues to evolve.
