The autonomous safety features are a bad idea. By trying to detect objects and stop, they introduce several problems: 1) The system has to successfully detect

The autonomous safety features are a bad idea. By trying to detect objects and stop, they introduce several problems: 1) The system has to successfully detect an object. If this system fails, it will still run over whatever was in the way. 2) The system has to successfully ignore detections which were false positives. When this system fails, the machine will just stop working, which may require a remediation which may cost more than the labor needed to pilot the vehicle. 3) The assumption that the vehicle is safe to operate autonomously because it has safety features will create a lack of concern for safety around the vehicle. This lowered level of concern will result in an increased number of incidents when the safety systems fail, as people will have been operating under the assumption that they won't fail. 4) In the worst case, if the safety systems fail often, they will be disabled (if possible), leading people to believe the system might still be somewhat safe, but unbeknownst to them it's completely unsafe. It would actually be safer to have no safety systems at all. That way everyone has to assume that the vehicle is unsafe, and therefore they will need to implement their own safety procedures, such as closing off the field it's being used in and not allowing anyone nearby into the field. This conscious fear of the device, combined with a conscious attempt to secure the area, will keep people safer.