Catalyst Game Labs > Errata

AI Quality "Errata"

(1/3) > >>

One major issue we're seeing with AIs on various communities is that as the release of further books has continued since Data Trails, AI have never received any updates on the Qualities they have access to. There were already some puzzling decisions from DT with AI Qualities (such as not being able to take the Quick Config quality) but this is exacerbated as time goes on, and AIs within the roles they're capable of fulfilling slide further and further behind metahuman characters who can use the full array of qualities and the various combinations that they create. This has become a more glaring issue with Kill Code, which is a book loaded to the brim with poorly-balanced, high-powered stuff that AIs don't have access to, along with stuff like Instinctive Hack out of BtB.

A useful post-DT "errata" for AIs would be to review the various books besides the five that include AI-enabled qualities (Core, Run Faster, R&G, Data Trails, Assassin's Primer) and determine qualities that make sense for AIs to take and aren't game breaking. There's probably at least one in most books, and at least a few in more technology-focused books. Having a comprehensive list would help get Chummer5/Herolab support for it (I don't know if Lab blocks AIs from taking non-AI qualities, never used it, assume it does), as well as making life easier on GMs instead of having to ad-hoc decide every time a player is interested in a quality from dozens of other sourcebooks whether it's AI applicable.

Another really useful QOL improvement would be for new sourcebooks to possibly have a little icon by quality names denoting if they're AI-enabled, or even just a (AI) supertext.

What say?

AI Qualities From Chrome Flesh

As an example, figured I'd look through a book's quality list and suggest some qualities that should be AI-applicable. Feel free to tell me I'm wrong.


Better to Be Feared Than Loved - AIs can certainly have exploitative relationships and many people are frightened of them (but would happily turn on them, given a chance to do so without reprisal)


Dry Addict - AIs can become Addicted (though I've no clue what to...BTLs?) via Fragmentation, so they should be able to dry out.

Lack of Focus - AIs have a lot on their mind and stuff they probably want to devote their processing cycles to. Should be mutually exclusive with Multiprocessing.

Poor Self Control (Sadistic) - The trope of the amoral AI who loves to experiment on and cause harm to other creatures.

Superhuman Psychosis - Sigh. I hate this quality but it's another kind of obvious one for amoral AIs who see others as lesser lifeforms.

Tough and Targeted - For an AI who becomes known, particularly around the Matrix (but also possibly as a rigger or jumped-in drone samurai), it makes sense that opposing hackers would love to take the Superbrain down a peg. Or maybe they're just seen as this untouchable hacker without even knowing what they really are.

A.I.s need a complete rewrite.

They were not functional at the time of publishing and as you have noted have only gotten worse since.

Unfortunately that sort of complete rewrite is out of scope for the errata team.

Which means A.I.s are going to continue to suck hard until they (maybe?) get rewritten by someone competent in the next edition of shadowrun.

TL:DR don't expect help from the Catalyst on this one, nor the errata process. Fix it yourself with your own home rules.

Honestly, you’re almost better off building a standard Metahuman Decker or Technomancer and just never having them come to the meet or on the ‘run in person. You’d end up with a character that “feels” like an AI, but has significant mechanical advantages... which is a damn shame.

I love the idea of playing an AI, but the rules for them just don’t quite work. I know it’s probably not economical from a publishing standpoint, but I really feel AI’s need their own dedicated splatbook that’s just about how to build one, how to role play one, and how to “roll play” one.

Spinx did some work on that here:


[0] Message Index

[#] Next page

Go to full version