Seed: Omniscient Killer AI

What if characters pissed off an artificial intelligence that has full access to hack the internet of things? Here’s some guidelines fit for a modern day or cyberpunk kind of game.

At the start of the adventure, the game master assigns each character a secret rating in “Awareness.”

Each time a character does some sort of research about the AI – internet search terms, business visits to technical experts, purchases high-end processing equipment – the GM increases “Awareness” by 1. Once per hour, the GM rolls a 1d6. If the result is under the “Awareness” number, the AI locks onto that characters as a potential threat to be tracked and – if necessary – eliminated. If the die roll is equal to or greater than the “Awareness value,” the AI has not yet noticed the characters’ investigation.

So what happens next when the AI seek to remove the character?

* Call the character’s phone and issue a “last warning” to drop the investigation or else.

* Hack into spy satellites and CCTV security systems to watch the character’s movements.

* Monitor any use of IDs to track characters – credit card transactions, cell phone calls, building ID keys, transit passes and so forth.

* Listen in on any cellular or electronic communications to or from the character, then reply with a perfect voice impersonation of friends, bosses or co-workers – anyone who could feed misinformation or lure the character into an ambush.

* Drones, drones, drones, drones, drones.

* Add the characters to the TSA No-Fly list or other such security watch lists.

* Wipe out any financial accounts connected to the character (checking, saving, stock trades, credit cards, property ownership records).

* Issues APBs (all-points bulletins) to all law enforcement agencies to capture the character at once. (Add “Use extreme caution, suspect armed and dangerous, has killed police in prior incident” to maximize prejudice against the target.)

* Seed Facebook shares and other social media with false news reports identifying the character as a terrorist, plague carrier, sex offender or worse – anything to cripple social connections.

* If the character travels in a modern vehicle with any kind of remote electronic control (new model cars, subway trains), the AI will hijack controls and cause the vehicle to crash.

* If the character lingers in a modern office building, set off sprinkler systems, raise or lower HVAC temperatures, lock doors, turn off lights and blast sirens or loud music over building announcement speaker systems.

* Order local sheriffs’ offices to foreclose on any properties being used by the character. Seize their personal equipment and put it all up for auction/liquidation.

* Falsify the character’s medical records to add conditions and diagnoses requiring dangerous medicines to treat. (Examples: Methotrexate for “ongoing cancer treatment,” Coumadin or Warfarin blood thinners to turn the smallest cuts into never-ending bleeds.)

* Anonymously hire hit-men, mercenaries, bounty hunters or other guns-for-hire to go after the character. Another tactic along same lines: Send through paperwork to get dangerous criminals released from prison in exchange for such killers agreeing to hunt down and kill the character.

* If the character cannot be precisely found but their general location is known, issue an emergency or set off alarms to scare the public, incite riots and spread disorder to “flush out” the character.

* If the above scare tactics don’t work, cause the in-flight systems of a passing jetliner to fail in such a way that the aircraft crashes on the character’s last known location.

And last but not least…

* Send a killer android built to pass for human out to track down and kill the target (aka “give ’em the ‘ol Sarah Connor treatment”).