robots:HCOM

Humans will have to adapt to the special needs of robot assistants.

Droids should repeat what they think you want before they store it on task or leave your vicinity, using an algorithm for restating things while saying them towards the "highest-likelihood wrong" while technically correct.

A typical conversation might go like: H: Get me a drink from the bar. A: I don't know the word "bar". H: Get me a drink from the saloon. A: I don't know the word "saloon". H: Say the words you have for "places to drink". A: "Fountain", "_",... H: That's the one. A: Which is "the one"? H: The last one. A: You would like a drink from _? H: Yes. A: The _ lists 42 drink on tab. Would you like one of these? H: Yes, the milken sweater. A:

Humans could consider communicating to droids using a musical flute to imitate the droid's robots:RCOM.

---

Bayesian networks are used to augment android calculations of risk and trust. The owner sets his or her own trust value (0.9) to anchor a baseline. All other trust scores are below this one, lest the droid become subordinate and change loyalties. Since it is not the goal of this manufacturer to create a world of droids to replace humans, it is the task of humans to become smart enough to train their droids and other droids that they come into contact with. So (like Advogato.org), an initial trust value of the owner is set, from [0.0,]1.0), and then other trusts in this range are multiplied wrt to this.