AIs & Muses
AIs are self-aware software. Like other code, they must be run on a computerized device. They may be crashed in mesh combat, copied, erased, stored as inert data, infected with viruses, and reprogrammed.
ALIs
Artificial limited intelligences are designed as “custodian helpers” for specific devices or tasks. They are based upon machine-learning neural nets that draw upon vast data sets concerning the device/operations, meaning that they are well-versed in the best methods and likely outcomes. They are programmed with the core skills they need for their tasks and equipped with a personality expressly geared towards customer service, obedience, and satisfaction.
Most ALIs are run on bots, vehicles, and other devices to assist transhumans or operate the machine themselves. Sample ALIs are listed on ALIs and Muses.
In most societies, ALIs are considered things or property rather than people and accorded no special rights. They may also be banned from certain roles or activities. A few polities treat them as persons with limited or full civil rights and protections.
Personality
ALI personality matrices are encoded with individual identities and character traits. They seem conscious and self-aware, though detailed probing will reveal shallow personalities and other limitations. They lack self-interest and initiative, and their emotional programming is narrow (based on empathy alone) or non-existent, though they may be programmed to anticipate the needs and desires of users and pre-emptively take action on their behalf. Though sentient, it remains a matter of debate whether ALIs are fully sapient.
Given their limited personality, ALIs are less emotive and more difficult to read. Apply a −30 modifier to Kinesics Tests made against them. When combined with non-expressive synthetic morphs, increase this modifier to −60. Some ALIs lack emotive capability altogether and are simply impossible to read.
Commanding ALIs
Due to built-in safety features, ALIs must serve and obey the instructions of authorized users within their normal functioning parameters. They are also programmed to follow the law (in the inner system) or an ethical code (autonomist areas). If need be, however, ALIs can be quite clever in how they interpret commands and act upon them. More importantly, their psychological programming is based on human modes of thought and an understanding and support of transhuman goals and interests. This is part of an effort to create “friendly AI.” An ALI will be reluctant to follow commands that will have a negative impact on the user or other transhumans. Pre-programmed imperatives can force it to ignore or disobey orders that are dangerous. In the case of accidents, they are sometimes coded to minimize harm and damage and sometimes coded to protect the user at all costs. Of course, ALIs designed for military and weapon systems do not face such restrictions.
Limitations
Lacking generalized intelligence, ALIs are inept and clueless when it comes to situations outside their specialty. Secondary skills tend to be less competent than an equivalent transhuman. They are generally incapable of creative or “outside-the-box” thinking.
Aptitudes, Skills, and Pools
ALIs usually have aptitudes of 10, with a maximum of 15. Most have 5–7 active skills, with a maximum of 10. Their skills cannot exceed 40 (including aptitudes), with the exception of one active skill appropriate to the ALI’s specialty, which can be 60. Most have 3–5 Know skills, with a maximum of 10. Their Know skills cannot exceed 80.
ALIs cannot default; if they don’t possess a skill, they can’t use it. They can use specializations (which can exceed their maximums).
ALIs do not get pools, unless a morph, bot, vehicle, or other device they are operating provides them.
By design, ALIs are incapable of self-improvement. To acquire or improve a skill, they must be reprogrammed or use skillware. They do not earn Rez Points.
For all other rules, ALIs are treated as infomorphs.
Traits
ALIs have the Enhanced Behavior (Obedient, Level 3) and Real-World Naiveté traits. GMs may assign other traits as they see fit.
Lucidity and Trauma
ALIs are capable of suffering mental stress and trauma, and so have Lucidity, Trauma Threshold, and Insanity Rating stats.
Sleeving
ALIs use their own individual infomorphs. ALIs may be sleeved into any morph with a cyberbrain. They cannot be downloaded into biomorphs with biological brains.
Muses
Muses are a subtype of ALI designed to be personal aides and companions. Most transhumans grew up with a muse at their virtual side. Muses have more personality and psychological programming than standard ALIs and over time they build up an extensive database of their user’s preferences, likes and dislikes, and personality quirks so that they may more effectively be of service and anticipate needs. Most muses reside within their owner’s mesh inserts or ecto, where they can manage their owner’s PAN, communications, online searches, rep interactions, and other mesh activity.
Personality and Relationships
Muse personalities are customizable. They are bundled with a subscription to a library of traits, quirks, and affects, many based on historical or fictional personas. Muses may also be modeled on personality templates of actual people; some are built from life-logged interactions with relatives or friends, others are based on fictional characters, childhood pets, or imaginary friends. While your first muse as a child was molded to suit you according to a battery of tests, over time you altered it to fit your changing tastes.
Transhumans develop strong bonds to their muses due to their omnipresence and devotion. Your specific relationship may vary, depending on your own personality, history, and views on sentient programs. Some transhumans treat their muses as intelligent toys, servants, or pets. Others are closer, viewing them as confidantes, mentors, comrades, or even paramours. A few, however, have contentious or even hostile relationships, especially if the muse has taken on an unwelcome guardian role.
What Your Muse Can Do for You
The primary use for muses is to handle trivial online tasks, thus freeing you up for more important things. This can be quite handy during time crunches or chaotic situations — it is otherwise quite inconvenient when you need to google something in the middle of a firefight. If you are not skilled in Infosec and don’t have or trust a team hacker, your muse can also act as your PAN’s system defender. Its important to remember that muses can also assist you in certain tasks, providing a teamwork bonus.
Here are a few specific examples of tasks your muse can take on:
- Protect your PAN as system defender.
- Shield your account shell or other software in mesh combat.
- Make Research Tests to find information for you.
- Falsify or fluctuate your mesh ID.
- Scan newsfeeds and mesh updates for keyword alerts.
- Teleoperate and command robots and ALIs.
- Launch countermeasures against intruders.
- Monitor your rep scores and alert you to drastic changes.
- Automatically ping and ding other people’s rep on your behalf.
- Run audio through an online, real-time language translation service.
- Put you in privacy mode and proactively stealth your wireless signal.
- Track people for you.
- Anticipate your needs and pre-empt your requests.
- Monitor your health (via medichines) and the status of ware/gear.
- Summon help if something happens to you.
Online, Even When You’re Down
Muses do not sleep, remaining active while you rest or hibernate.
Likewise, if you are incapacitated or grappled, your muse can still act (Damage and Infomorph Riders). Shock attacks, however, temporarily disrupt implanted systems, so your muse will be incapacitated just as you are.
Muse Backups
Muses are normally backed up and stored along with the ego, including on cortical stacks.
Other Infomorphs As Muses
Some people prefer full-blown intelligences for aid and companionship over a muse. Any infomorph can take a muse’s residence within your mesh inserts, whether an AGI, disembodied transhuman, or even a fork of yourself.
Roleplaying Muses
Muses provide boundless roleplaying opportunities. GMs can use them to remind, nag, encourage, or pester PCs — no one is ever really alone. If you find it overwhelming to keep track of the PCs’ muses and their personalities, have the players roleplay each others’ muses. Simply give each player an index card with the muse’s skills and personality notes; hold the card up when roleplaying as the muse. This can also be a useful tool for keeping players involved in a scene when the party splits up and their own PCs are busy elsewhere.
AGIs
AGIs are complete and fully autonomous digital consciousnesses, self aware and capable of intelligent action at the same level as other transhumans. AGIs are capable of creativity, learning, and self-improvement (at a slow but steady pace equivalent to humans). They also possess deep, rounded personalities and stronger emotional/empathic abilities than standard ALIs. Most are raised in VR in a manner similar to human children, so that they are socialized with transhuman values. As a result, they have a fairly human persona and outlook, though some deviation is to be expected — AGIs often possess or develop personality traits and idiosyncrasies that are quite different from human norms and sometimes outright alien. Despite this attempt to humanize AGIs, they do not have the same evolutionary and biological origins that transhumans have, and so their social responses, behavior, and goals are sometimes decidedly different. On an emotional level, AGIs run subroutines that are comparable to biological human emotions. Most are programmed to have empathy, share an interest in transhuman affairs and prosperity, and place significant relevance on life of all kinds. In game terms, AGIs emote like other transhumans (and so Kinesics may be used to read them) and are vulnerable to emotionally manipulative effects, fear, etc.
AGI minds emulate transhuman neural patterns, allowing them to sleeve into morphs with biological brains. AGIs sometimes have trouble adapting to biological neurochemistry, finding reactions such as hunger and fear to be as confusing as they are novel.
AGIs bear the social stigma of their non-biological origin and are often met with bias and mistrust. Some habitats outlaw AGIs or subject them to strict restrictions, forcing such infolife to hide their true natures or illegally darkcast in.
Just like other PCs, AGIs earn Rez points and may improve their skills and capabilities. AGIs suffer none of the skill limitations placed on ALIs, using skills just like any other character.
ASIs
Super-intelligent AIs can self-upgrade at exponential rates and grow into god-like digital entities. They require massive processing power and are always increasing in complexity due to a continual metamorphosis of their code. The only ASIs publicly known to exist are the infamous TITANs who are widely regarded as being responsible for the Fall. Many suspect that the TITANs were not the first ASIs, however, and they may not be the last. No ASIs are currently known to exist within the Solar System or on any exoplanets visited by transhumans, though rumors circulate of damaged TITANs left behind on Earth, speculated TITAN activity under the clouds of Venus, or whispers of ASIs hidden away in secret networks on the edges of the system.
ASI minds are too vast and complex to be downloaded into a physical morph, even a synthetic one. They have been known to sometimes create massively dumbed-down forks that can sleeve into physical morphs, with drastically reduced mental capacities.
ASIs are too intelligent to codify in game terms. They can think circles around transhumans, have plans within plans within plans, have unlocked secrets of the universe that we only dream about, and are unlikely ever to be caught off-guard.
Non-Standard AIs
Not all ALIs and AGIs are programmed and designed to adhere to human modes of thought and interests. These include emergent neural networks that expressed intelligence on their own, uplifted expert systems, machine minds based on non-human brains, and stranger things. Such creations are illegal and considered a potential threat in many jurisdictions. Nevertheless, hypercorps and other groups experiment with varying results. In some cases, these digital minds are so different from human mindsets that communication is impossible. In others, enough crossover exists to allow limited communication, but such entities are invariably quite strange.