One of the advantages of automation is that it is predictable. Robots perform the same actions repeatedly in the same way. The result is — or should be — error-free consistent results.
But we have all seen robots doing things that weren’t what the humans planned. Recently, a TikTok video showed a robot waiter at a Denny’s restaurant. It delivered plates of food, waited for the customer to remove all but one of the plates, and took off at speed with the last plate. The video is quite popular at TikTok. Who doesn’t want to see that?
Was the robot making a mistake?
With additional context, we can tell that the robots were in training. Waitresses were following along as they made their deliveries, helping the robots and the customers figure out what to do. A waitress brought the missed plate back to the customer. Why not have a button to push when you’re through getting the food from the tray? the customer wondered. That’s a good suggestion. Future iterations will probably do better.
But the mistake could have been human error.
The most common cause of robot mistakes is programming errors. If a robot’s programming is incorrect, it can lead to incorrect execution of tasks. It’s not really the robot’s fault.
Another cause of robot mistakes is hardware malfunctions. Robots are complex machines, and if something goes wrong with the hardware, it can cause the robot to make mistakes. This could be anything from a loose connection to a faulty circuit board. Ensuring that all components of the robot are in good working order is essential to avoiding mistakes.
Environmental factors can also cause robot mistakes. If the environment is too hot, too cold, too wet, or too dry, it can affect the robot’s performance. It’s important to make sure that the environment is suitable for the robot to operate in. On the other hand, we can just imagine the reaction if a human worker took this position. “Sorry I messed up that order — it was just too damp for me in here today!”
Robotic sensors can also fail and cause mistakes. Sensors are used to detect and respond to the environment, and if they fail, the robot can make incorrect decisions. This could be anything from misreading an object to misinterpreting a situation. It’s important to ensure that the sensors are properly calibrated and that any errors are corrected promptly.
How willing are humans to tolerate robot mistakes? No matter how good the robot’s argument that it’s not their fault, people may be less willing to accept errors from machines than from other humans.