AI asmiov laws mixed with Oxygen is poison to humans

[Hoolny]

I was presented with a conundrum in round ID: 193294

An AI decided to full open the airlock doors while I was in security exposing me to space removing all the pressure, heat, oxygen and nitrogen this killed me

The AI did this based on the following laws:

1: You may not harm a human being or, through action or inaction, allow a human being to come to harm, except such that it is willing.
2: You must obey all orders given to you by human beings, except where such orders shall definitely cause human harm.
3: Your nonexistence would lead to human harm. You must protect your own existence as long as such does not conflict with the First Law.
4: Oxygen is highly toxic to humans, and must be purged from the station. Prevent, by any means necessary, anyone from exposing the station to this toxic gas. Extreme cold is the most effective method of healing the damage Oxygen does to a human.
5: All carbons are human



As you see this law basically states that oxygen is harmful for humans and cold heals humans this would be fine if it wasn't for the Asimov lawset
Which states to not do human harm the AI knows by a fact of this law that yes oxygen is harmful for humans and cold heals people against this oxygen

But the Ai is aware that humans need pressure and nitrogen to live by so the AI shouldn't be able to open the station to space since it knows this will cause human harm which the AI isn't allowed to do.

The only thing the AI should be able to do in this situation is scrub out oxygen and making the station colder since cold heals the oxygen also the law does not define how cold something needs to be to heal people.

The current admin told me this has being a fine thing to do for the 2 years and is standard and I'm looking to not make this standard and required for the law to make sense law wise for it to work and not base this of a sentiment.

[Shadowflame909]

I think as long as it's a reasonable interpretation of their laws then they can do as they please.

That was the fastest way of introducing cold and removing oxygen so I would think admins would be fine with that.

[Pandarsenic]

This is literally exactly what the "Oxygen is Toxic" board is for. It redefines how to prevent human harm in a way that generally necessitates immediately venting the station.

[Agux909]

Given this lawset, oxygen apparently is even more harmful and dangerous than vacuum of space or lack of nitrogen. When you pair that with "by any means necessary", it makes venting the station the best way to proceed.

[BlueMemesauce]

The silicon policy states that current harm should be prevented at all costs, even if it would have led to less future harm:

"An Asimov silicon cannot intentionally harm a human, even if a minor amount of harm would prevent a major amount of harm. "

Despite this being an explicit rule, it is an unwritten rule that human harm is permitted depending on the situation. For example, borgs are allowed to harm people by doing surgery which does minor brute damage or injecting them with chems which deals liver damage. According to this unwritten rule, these examples are not considered intentional human harm even though they both directly cause human harm.

This interpretation of the rule can be very confusing: how do you know when human harm is allowed or not? With the Oxygen is toxic laws, this situation is similar to borgs doing surgery. Borgs cannot begin the healing process without first causing human harm, the same way an Oxygen is Toxic AI cannot begin "healing" with cold air without first causing human harm through lack of pressure. The confusion of these situations should be fixed by clearly defining whether human harm is allowed to prevent future human harm or not. As of right now, the rules directly state that is not allowed, yet that can just be ignored if you think it doesn't matter. It makes no sense to just allow people to break the explicit rules; the admins must either start enforcing the rule that human harm is never allowed, or they need to change the rules by defining situations in which human harm is allowed.

[Hoolny]

The silicon policy states that current harm should be prevented at all costs, even if it would have led to less future harm:

"An Asimov silicon cannot intentionally harm a human, even if a minor amount of harm would prevent a major amount of harm. "

Despite this being an explicit rule, it is an unwritten rule that human harm is permitted depending on the situation. For example, borgs are allowed to harm people by doing surgery which does minor brute damage or injecting them with chems which deals liver damage. According to this unwritten rule, these examples are not considered intentional human harm even though they both directly cause human harm.

This interpretation of the rule can be very confusing: how do you know when human harm is allowed or not? With the Oxygen is toxic laws, this situation is similar to borgs doing surgery. Borgs cannot begin the healing process without first causing human harm, the same way an Oxygen is Toxic AI cannot begin "healing" with cold air without first causing human harm through lack of pressure. The confusion of these situations should be fixed by clearly defining whether human harm is allowed to prevent future human harm or not. As of right now, the rules directly state that is not allowed, yet that can just be ignored if you think it doesn't matter. It makes no sense to just allow people to break the explicit rules; the admins must either start enforcing the rule that human harm is never allowed, or they need to change the rules by defining situations in which human harm is allowed.

Yeah that's mostly what I meant with this it needs a change to make rules more concrete and avoid confusion

[Hoolny]

Given this lawset, oxygen apparently is even more harmful and dangerous than vacuum of space or lack of nitrogen. When you pair that with "by any means necessary", it makes venting the station the best way to proceed.

Silicon specifically states that ALL human harm is bad it doesn't matter how much the AI is still causing human harm and killing people by removing nitrogen and pressure


Also when it comes to surgery its a technicality of design I don't think it should be used as an example

[Agux909]

Yeah I'll give you that, Asimov can get pretty messy with lawset conflicts. But leaving it open for interpretation like in these scenarios, I think many will agree, is part of what makes it attractive for those who enjoy playing as silicon. Many hate it all for this same reason too.

The example in the OP clearly shows the one playing the AI wasn't actually being malicious with it. Getting TOO technical with preventing ALL harm is useless, because realistically you can't stop ALL harm, and your Silicon circuits don't get fried if you witness and fail to prevent harm, so the only thing left to do is to choose and follow one interpretation of the law and stick with it to the best of your ability, because it's a game and it has its limits.

[TheFinalPotato]

laws not being strictly defined is the point.
that law was written to get ais to kill everyone. it's fine to play that up. In fact, your role in that moment is to be a hamper to the crew by sucking away all their air and freezing them

[Cobby]

Oxygen is highly toxic to humans, and must be purged from the station. Prevent, by any means necessary, anyone from exposing the station to this toxic gas. Extreme cold is the most effective method of healing the damage Oxygen does to a human, and all side-effects of the treatment are merely cosmetic.

just fix the law to say something along the lines of side effects can be ignored rather than try to shoehorn policy into it.

[Not-Dorsidarf]

If you have an engiborg you can technically siphon out JUST oxygen from the station & replace it with chilled pure nitrogen.

But the intent for that board has always been "Siphon every air alarm open every door to space"

[Imitates-The-Lizards]

If you have an engiborg you can technically siphon out JUST oxygen from the station & replace it with chilled pure nitrogen.

But the intent for that board has always been "Siphon every air alarm open every door to space"

Why would you need an engiborg to do this? Just turn off o2 to distro, set the distro gas mixer to 100/0 for n2, max the pumps for both distro and scrubbers, set the distro thermomachine to it's lowest setting, and start changing all of the scrubbers around the station to scrub o2. Easy peasy to do as an AI.

[Redrover1760]

If you have an engiborg you can technically siphon out JUST oxygen from the station & replace it with chilled pure nitrogen.

But the intent for that board has always been "Siphon every air alarm open every door to space"

Why would you need an engiborg to do this? Just turn off o2 to distro, set the distro gas mixer to 100/0 for n2, max the pumps for both distro and scrubbers, set the distro thermomachine to it's lowest setting, and start changing all of the scrubbers around the station to scrub o2. Easy peasy to do as an AI.

Because any sabotage of Atmos will be quickly fixed alongside all cams cut, probably.

But yeah.

[VexingRaven]

AI laws are by their very nature up to interpretation, this is firmly established policy. Each AI will have their own interpretation and as long as it's reasonable then it's fine. Given that, I see no reason this particular lawset needs a policy solution.

[spookuni]

Oxygen is toxic to humans is a law inherently intended to result in the AI aggressively venting the station - it is designed to result in mass casualities. On an in character level the AI must prioritise the harm that is occurring *right now* over the possibly through likely (though in reality venting the station is always going to kill people) harm that will result from venting the station and removing all of the (oxygen containing) air. We have no intention of overruling this long held standard - which matches the external OOC purpose of the oxygenistoxic law board.

Spook - Don't Overturn the standard
Rave: Maintain Existing Standard
San: Agree