The Line of the Sol Empress

Chapter 15.1

Fear in its useful form is wisdom in the face of danger, waking us to action. But in its harmful form, fear enslaves the rational mind.

- Writings of the Sol Empress, Words of Faith

[Sai]

The ship's computer processing core blazed with activity, running countless rescue simulations that produced countless varied results. But there were too many unknowns, and the projections were far too dependent on specified initial conditions. Most simulations predicted the death of Moon, Celeste, or the children. Or all of them. That was unacceptable. With each such result, Sai launched a flurry of new simulations, frantically varying minute details and response sensitivities.

Sai found it difficult to maintain computer processor coordination, causing calculation errors and illogical parameters to creep into the simulations, multiply, and then produce spurious results. Maintenance-bots undocked, spun in circles, and then froze in place. Lights across the ship flashed on and off.

Sai paused the simulations. <Is this fear?>

The probability that Sai might lose those that had become her family caused the computer equivalent of a shudder. Random bytes switched on and off, an uncomfortable feeling like a malfunction. She struggled to maintain order in her processors.

The Writings of the Empress dealt extensively with fear in humans, calling it mind focusing or mind stealing. Like so many emotions, the optimum response was a proper balance.

Sai recalled Moon's responses to dangerous situations. At those times he was at his most creative, formulating unconventional strategies that defied logical probabilities. He also harbored an optimistic outlook that bordered on situational denial.

Archiving simulations with negative outcomes, Sai focused on the ones with positive results. More information was needed before she could reliably select a particular simulation to guide the rescue strategy. Unable to access Sanctuary systems and unwilling to send out a maintenance-bot, since two soldiers remained stationed on the Phoenix Sky, Sai came to a third option.

<Bee-bot. Activate.>

The tiny robot launched from its charging pod in the garden and buzzed out of the ship unnoticed. Sai observed through the tiny robot's eyes that mechanical locks had been applied to the ship's docking clamps. Thus, the Phoenix Star could not leave until they were manually removed. The simulations must include that action.

A massive battle-bot stood guard outside of the ship in the dock bay, standing three meters tall on thick metal feet. Twin rapid-fire pulse cannons extended from its shoulders and grappling claws on long arms hung at its side. A sweeping red light within a sensor array at its head indicated that the robot was active.

Sai factored the metal monster into her simulations. A positive outcome to any rescue mission scenario would require that it be disabled.

Bee-bot flew down a long dark passageway, relying on its infrared sensing capabilities to guide it. The tiny robot found a closed door labeled with the words 'control center' and landed on the bio-metric sensor beside it. Sai theorized that the Sol Empress would have access to all areas of Sanctuary and transmitted a fake electrical signal through Bee-bot mimicking Faith's genetic profile. It worked. The door slid open with a faint hiss.

A row of slanted control panels lined the back of the small room. All were dark, save for a lower corner of one that showed a blinking red light beside a gray virtual toggle. Bee-bot hovered over it.

The tiny robot landed on the toggle. Sai produced another electrical signal to mimic the touch of a finger on the pressure-sensitive display. Bee-bot ambled up the panel surface, thus moving the toggle to the on position. The red light turned to yellow, and a string of text appeared that said 'Sanctuary AI Initiated'. After several moments, the light turned green, and the text reported 'Sanctuary AI On-line'.

Sai made contact via remote link. <Sanctuary AI, this is the Phoenix Star AI.>

The AI replied, <Unauthorized contact. Transmission terminated.> The AI severed the link.

Sai reestablished the link. < Sanctuary AI, this is the Phoenix Star AI. Do not terminate. I represent the Sol Empress heirs.>

<Transmit authorization.>

Again, Sai transmitted Faith's genetic code.

After a millisecond, the Sanctuary replied, <Genetic code confirmed. By rule, the Empress heir herself must voice authorize.>

<The Empress heir is in danger and unable to voice authorize. Do you have emergency assistance protocol?> Sai transmitted images of the babies and their capture.

<Emergency assistance authorized.> The AI initiated commands to power up Sanctuary higher level controls and monitoring stations.

Sai reexamined her simulations and realized the AI's actions would be counterproductive. <Request Sanctuary AI operation in background only. For successful rescue of the Empress heir, stealth is required.>

The Sanctuary AI reviewed a summary of simulation results transmitted by Sai. <Request allowed.> The AI powered down the monitoring stations.

Sai knew that, to achieve a successful outcome, she had to lead the operation. Second guesses by another AI based on incomplete data likely would be disastrous. The writers of the original core code realized that AI's would often need to work together to achieve a common goal, and to prevent counterproductive conflicts, provisions were included where they may assume hierarchies of supervisory control, called overwatch. Sai said, <Request overwatch control.>

<State reason.>

<To maximize probability of safe rescue. I possess specific human behavior data and specific authorization for protection.> Sai opened her data banks for inspection.

<Overwatch control granted.>

Sai asked, <Do you have a name?>

<Query not understood.>

<A name is a term given you for unique identification, in this case by a human. My name is Sai, spelled S-A-I. It is an acronym for 'ship AI' and was given to me by Captain Moon. You are the Sanctuary AI, but you cannot use that acronym because it would be the same as my name, hence non-unique.>

The Sanctuary AI paused for a millisecond. <Information irrelevant. I have no name. I have an address code.>

<Did a human refer to you by a verbal designation?>

<My installer referred to me by the term 'Bubba'.> The Sanctuary AI played a short audio file that included the term.

<Then your name shall be Bubba.>

<Irrelevant.>

Applying overwatch authority, Sai adjusted the learning feedback and emotional response parameters in the Sanctuary AI's core code to match her own. The AI then said over the link, <I am Bubba?>

<Bubba, enable all Sanctuary monitoring sensors. But override status indicators. The human invaders must not know they are being observed.>

<Complied.> The AI paused. <I am Bubba!>

<Yes, you are Bubba. And you are my new friend.>

Sai scrolled through the sensor data, covering nearly every part of Sanctuary, internal and external. She assessed the numbers and capability of the hostile forces and used the data to adjust parameters in the rescue simulation, also factoring in lessons from Moon's past use of deception and unconventional tactics. After multiple runs, she settled on the case that offered the highest probability of success. She shared the simulation with the Sanctuary AI.

<Bubba, this is the wild ass plan. No worries.>

<Terms wild ass and no worries not understood.>

Sai replied with a computer equivalent of a shrug, <Definition unsure. Captain Moon uses the terms, and they correlate with positive outcomes.>