AI Ethics
How AI is allowed to behave across Rover Planet, in plain terms.
Safety first
Our AI refuses requests that could cause physical, psychological or financial harm. When something looks unsafe, it stops and asks a person.
Respect for everyone
No harassment, no demeaning content, and no discrimination on race, gender, religion, orientation or any other characteristic.
Honest about being AI
It never pretends to be human, it says when it is uncertain, and it is upfront about what it cannot do.
Your data stays yours
We collect only what a feature needs, encrypt it in transit and at rest, and never share it without your consent. You can delete it at any time.
People make the big calls
AI recommends; people decide. Consequential actions — money, access, anything hard to undo — need a human, and every automated action is logged.
Fair by design
We test our models for biased outcomes across groups and fix what we find, rather than assuming the results are even.
Genuinely useful
AI should make you more capable, not replace your judgement. If a use case only works by deceiving or exploiting someone, we do not build it.
How we hold ourselves to this
AI-generated content is screened automatically, consequential decisions route to a person, and automated actions are logged so they can be traced back and explained. These principles change as the technology does — when we get something wrong, we would rather hear about it and fix it.
Seen our AI behave in a way that contradicts any of this? Tell us at ethics@roverplanet.com.