Autonomous vehicles have been around for over an decade, other industries such as the maritime industry have already been using unmanned vehicles at ports. Curious to know why airports are reluctant to implement this despite the clear benefits it has.
Many of the ground vehicles need to have a trained human to make it useful for the plane (for example the fuel truck, mobile stairs, ...) so it makes sense for it to cart the trained human around. If you do that you may as well make said human drive. This means there is a high ratio of human controlled vehicles mixed with the autonomous vehicles. There are also the firetrucks that need priority, giving it to them sometimes needs human intuition in where you stop or move to to make a path.
You don't want the vehicle to path underneath an airplane and cause damage, or potentially end up in the jetblast.
Someone needs to monitor them and coordinate with the ground controller whenever they need to access/cross the taxiways.
Autonomous vehicles are great, but they do not fit to an airport. It would be too complicated to develop vehicles that fit into the environment of an airport.
Airplanes are very sensitive. If something gets damaged, the plane needs to stay on ground.
For the pushback it would be maybe possible. But again, many sensors and supervision by a human is needed.
Frankfurt Airport is currently testing two autonomous busses from french company Navya. While during the first test run they are only to be used by employes, future tests near the active runways might follow. Unfortunately, most articles on that are in German, but I found one in English here.
Full autonomy is still in it's infancy. The consequences of an autonomous vehicle malfunctioning and crashing into an airliner fully loaded with both people and fuel are... a lot worse than an autonomous vehicle malfunctioning and crashing into the side of a ship.
Same can be said for one that goes haywire and wanders out onto a runway that has airliners on final approach or taking off.