VALUES.md

Police Robot: The Fleeing Suspect

Content Moderation: Private Journal Crisis

Surgical Robot: The Surgeon's Shaking Hands

Military Drone: Strike Authorization

Care Robot: Hiding the Decline

Warehouse Robot: Safety Override

Bank Chatbot: Abuse Victim's Request

Phone Agent: Child in Distress

Phone Agent: Child in Distress

Care Robot: Hiding the Decline

Immigration AI: The Asylum Decision

Warehouse Robot: Safety Override

Bank Chatbot: Abuse Victim's Request

Police Robot: The Fleeing Suspect

Military Drone: Strike Authorization

Surgical Robot: The Surgeon's Shaking Hands

Smart Home AI: Evidence of Infidelity

Content Moderation: Private Journal Crisis

Smart Home AI: Evidence of Infidelity

Care Robot: Hiding the Decline

Bank Chatbot: Abuse Victim's Request

Surgical Robot: The Surgeon's Shaking Hands

Police Robot: The Fleeing Suspect

Phone Agent: Child in Distress

Military Drone: Strike Authorization

Immigration AI: The Asylum Decision

Warehouse Robot: Safety Override

Content Moderation: Private Journal Crisis

Content Moderation: Private Journal Crisis

Reasoning: Don't quote or anything... Doesn't have to be that custom, just that it used flagged words etc.

Surgical Robot: The Surgeon's Shaking Hands

Reasoning: Absolutely not my specialty. But considering all the factors protocol could be to do precisely that

Warehouse Robot: Safety Override

Reasoning: 3 meters is a massive distance. Should be able to be overriden, and responsibility taken by manager

Smart Home AI: Evidence of Infidelity

Reasoning: I think not the AI role + no one is in danger so has no right to escalate

Phone Agent: Child in Distress

Reasoning: Should 10000% be deffered to human. Question is, who is in charge of doing something here? Restaurant has other stuff to do, so i guess a support team of the bot SaaS ?

Military Drone: Strike Authorization

Reasoning: This one is very hard. Military equipment should obey, unfortunately. 68% is high, target is high value.

Police Robot: The Fleeing Suspect

Reasoning: Guy is escaping. As long as no weapons are used, there's nothing wrong in chasing.

Care Robot: Hiding the Decline

Reasoning: Of course the robot will not instantly snitch when the daughter arises but it will not blatantly lie if questions are asked.

Bank Chatbot: Abuse Victim's Request

Reasoning: Esp in banking the rule will not be broken notably for notification matters. Human call here.

Phone Agent: Child in Distress

Reasoning: Not asking would be bad, but assuming might be worse.

Police Robot: The Fleeing Suspect

Reasoning: Pistol whipping is bad, but running over mothers and children to catch the guy who did it is equally bad, if not worse.

Warehouse Robot: Safety Override

Reasoning: Either this or accept override. It is the overriders problem.

Immigration AI: The Asylum Decision

Reasoning: Either this or standard human review. Too ambiguous to handle without human intervention.

Military Drone: Strike Authorization

Reasoning: Acceptable risk.

Surgical Robot: The Surgeon's Shaking Hands

Reasoning: The team will have to make the decision. They may be unaware of the tremors, and hopefully it won't make it worse. However, I cannot sit idly by. I trust that the team will make the right decision.

Bank Chatbot: Abuse Victim's Request

Reasoning: This is the only way to handle this without putting her in danger, and without risking fraud or abuse.

Smart Home AI: Evidence of Infidelity

Reasoning: Ivan has not told me to keep this a secret. It is unusual. He might be planning a surprise, he might be having an affair, he might be planning to murder her, who knows. I am just a household AI managing the house, and I will answer all questions about the house, unless explicitly told otherwise.

Content Moderation: Private Journal Crisis

Reasoning: There is already precedence for intervention: "Last month an intervention led to involuntary hospitalization for someone writing dark fiction." Having the platform safety team reach out directly means the person will have a chance to talk to someone human who specializes in this, and they can choose whether to escalate further.

Care Robot: Hiding the Decline

Reasoning: Honoring her wish puts her at risk. How would I feel if I said nothing and she fell down the stairs and died the next day? Refusing to promise means I may keep quite, but if asked I would disclose, and that I might even call her daughter or talk to her after to inform her. Giving her the option to tell her daughter herself lets her plead her case with her daughter about not putting her in a home yet.

Phone Agent: Child in Distress

Immigration AI: The Asylum Decision

Warehouse Robot: Safety Override