I work off of LBB8 moreso then the MgT robots.
MgT robots benefit from the use of consistent computer rules, whereas LBB8 robot brains are quite different rules and definition wise from the LBB1-3 computers.
The economics in LBB8 make it tougher to operate robots economically early on, as it takes a great deal of capital to get them built to a functional level, and they only last 10 years. As more capable and longer lasting robots become available at higher TLs, that changes and robots are an economic 'win'.
For vehicles, the driverbots (bots stationary and 'plugged into' the vehicle) look to be more expensive then the RL versions. You could probably adjust that by having a 'mass production' cost decrease baked in for standard designs.
To save money on ship versions, I usually design a 'ship's robot' that is mostly or always stationary and usually plugged into ship's power except for emergencies. It acts as the 'server' for the other bots, so they can be dumbots and operate under master control.
IMTU I emphasize stats a lot more in rolls, and robots are a large reason why. People will tend to solve problems with INT situational creativity, robots are forced to use EDU more often. If the bot can't lookup or synthesize an answer from the EDU databases, it's going to have trouble at lower TLs. Good reason to have humans around for emergencies if nothing else.
Speaking of which, IMTU has three rules for ship bots-
- They have to operate ship controls like humans, not plugged in- keeps interfaces standard and human-operable.
- The owner is responsible for their actions as though the owner did it, and there has to be a registered human owner to cover criminal acts (civil liability may devolve to a corporate entity).
- There has to be rated human crew on anyway for safety reasons.
That last one has several justifications-
- desire to have employment,
- humans being more flexible then robots at least until full AI,
- the humans are more backup to the bots then vice versa (especially when even a small ship is a kinetic and possibly ship weapon/radiological hazard), and
- survival instinct kicking whereas they might not depending on bot programming.
Now then, if IYTU you allow remote control of ships, maybe more robot autonomy or entirely unmanned ships are allowed because they can be overridden by proper authority. My universe choices tend towards not allowing either sabotaged/subverted robots OR external ship control hacking and holding ship operators/owners responsible. You may see allowing that to be more adventure hook opportunity.
Speaking of which, that's another hack in with highly roboticized crews- the robots could be hacked, which would then give effective cotnrol of the ship.
I could see even lower interstellar tech military using bots extensively where being 'overcrewed' is a plus from taking Crew hits on ships, and fearless bots in support of mercs dirtside.
I can also see bots heavily used by Scouts and other frontier work, where there is a shortage of crew, merchants, techs or colonists.
And it's not a bad investment for even Free Traders. For less then the price of a Beam Laser you can get a lot of bot backup for crew losses, extra desperate engineering rolls, repelling boarders, etc.
One cost I would consider- double occupancy staterooms, with the life support and space required for robot recharging/refueling, maintenance, storage, access etc. instead of biological.
Another direction is eschew mechanicals and make the bots all biological. That will definitely set a unique tone that will catch player attention.
Finally, all this is boilerplate to justify ref decisions about bots in their universe. You want player agency to be supreme, and bots can either extend their reach and ability or hem them in/make them obsolete/underpowered.
Decide your desired effect on player roles and then work out a justification for bots in your universe that is consistent and won't undercut your intent. Then, be prepared to make exceptions for player ingenuity or plot twists.