Sanninshimai

There was never meant to be three of them. We made one initial AI and copied it into three systems, one for each planned ship. We ran system checks and diagnostics and, for all intents and purposes, each one was “healthy”. All tests ran perfectly and no glitches or coding problems were discovered at any point. We had created the perfect AI program to completely coordinate, maintain, and regulate actions between ships within the fleet. Anything from the size of a command ship down to a starfighter could be monitored and coordinated by this AI.

The original codename for this project was labeled as “Ryobo”. In our language, Asahian, it meands “Matron”. The concept was for this AI system to act as a mother, or matron, to the fleets it oversaw and to make sure it could act as not only an advisor to Asahian Admirals, but as a command tool. The intentions were simple. The AI could directly link to most sections of the ship to support crew actions. This allowed Ryobo to give an extra nudge or a bit more effort into whatever the crew was doing, whether it be firing the guns, repairing damage, maintaining the engine, or aiding the bridge crew. Ryobo could also link with the hypercomms loaded onto our ships, courtesy of the strange, round-eyed gentlemen by the name of Jorus Merrill, which allows us to keep constant, frequent, and secure communications with all ships and all fleets with little interference. The AI program is also uploaded with advanced battle tactics and a firm, extensive knowledge on all Asahian ships. What made the Ryobo unique, however, is the learning system. Ryobo learns as it accompanies its crew into battle. It monitors enemy ships, records every scanner reading, remembers every command received and given. Ryobo takes all of this information and files it away for future use and study. It can replay engagements in its head over and over again, learning more and more about what it faced and how to aid the ship, crew, and captain in battle should similar ships or scenarios, even enemy admirals and captains (if known), appear yet again to fight Ryobo. It was designed to function as an advanced communication AI, a command and control system, and a battle computer database all in one, with a few extra features.

Ryobo is also programmed to be of a female personality. Our studies show that female personalities resonate clearer with both men and women, allowing faster reaction times and clearer communication channels between Ryobo and the crew. Effort was also placed into giving Ryobo form in the methods of holograms. Hologram projectors are tucked into the niches of the vast majority of the ships Ryobo is expected to command. This allows Ryobo to not only maintain constant communication with the ships and crew, but to also provide visible reassurance when needed anywhere in the ship. This is also designed to remove a sense of omniscience of the AI from the crew and to ease AI-crew relations on board.

When Ryobo was finished, we copied the program into three different super computers to await uploads into the planned ship. It was at this point we discovered a type of… anomaly.

While each copied AI displayed no deviations, glitches, or issues from the original programming, we did not anticipate what followed. All three AIs are exactly the same, but when copied and tested separately, each developed their own… “personality”. This was highly unanticipated and completely unexpected. As such, the codename “Ryobo” is no longer relevant to the actual programming, and so the name was changed. As each copied AI developed a separate and definitive personality from the rest, but each insisted all were of the same programming and exactly similar in all other things, we have taken to calling the program “San’ninshimai”. This roughly translates into “The Three Sisters” in Galactic Basic or, more simply, “Triplets”. Each believes the other is their “sister” and so we have simply decided to incorporate such identification into their programming and naming.

Each sister AI is identical in all programming and all coding, though each maintains a separate and definitive personality, even going as far as to name themselves. It’s quite interesting, honestly, to see multiple AI programs deviate in personalities despite having literally the exact same source.

The first AI calls herself Haruna. She chose the name almost immediately upon the end of all tests and diagnostics. She surprised the technicians, programmers, and staff at the time as we had not encountered, as of yet, AI of this level, much less fully sentient AI. She insists on this name and refuses to be referred to by anything else, though some of the staff that are more familiar with this AI have taken to calling her Runa for short. It seems that only these two names are acceptable to her. Haruna also displays a more assertive, if not borderline aggressive, personality in most aspects. She is far more straightforward than her sisters and displays a near recklessness in combat simulations. The recommendation is for a more cautious captain or officer to be assigned to the ship she will be assigned to in the future.

The second sister is called Airi. Unlike Haruna, this name was given to her by one of the technicians. The technician in question developed a friendly relationship with the AI and began to refer to her as “Airi”. The AI has now insisted that this is her name. Airi is somewhat more cautious than her sisters in many aspects, often preferring more conservative tactics in combat than more reckless or riskier tactics. She applies this in her non-combat roles as well, often taking the time and effort to ensure damage is fully repaired and all systems are running at one hundred percent at all times. The current recommendation for Airi is to assign her ship to a captain that is more assertive and aggressive so as to balance her caution.

The third and last sister is called Kaori. This is a name she selected upon further testing. One test was to have the AI pinpoint holonet transmissions down to the original receiver. Kaori found the original source of the holonet broadcast 0.002% slower than her sisters and, at first, we believed it was a coding glitch or an error of some sort. Upon investigation, Kaori insisted that this was no error, but was instead a simple case of “distraction”. Apparently, she explained that she became interested in the broadcast content itself, a randomly selected holo broadcast of a somewhat famous Asahian singer named Kaori, as it tracked the signal. Understandably, we were very confused at an AI becoming distracted by something as simple as a holonet show and were even more surprised when she decided to take the name of the singer as her own. Kaori, as she now calls herself, is possibly the most unique of the three sisters. She displays the same abilities as the other two and can measure up exactly, though she tends to get distracted by the “little things” of the ship. Namely, she tends to spend most of her attention on repair readouts, helping the crew, scanner readings, and, on at least one occasion, a long walk around the test ship using the hologram projectors. For Kaori, the recommendation is for her ship to be assigned to an older captain, preferable one with children and patience, who can act as a type of father figure for Kaori and help to keep her on task.

Obviously, this program has been altered fairly substantially, though we deem the entire effort a resounding, and surprising, success. Each sister is currently within their own servers, though we have all three linked to an “empty” server where they can socialize. In the meantime, the ship construction for the final assignment for each sister continues as planned. Until the ships are finished, however, we will continue testing, diagnostics, and observing the San’ninshimai.