(This is the significant beginning of a story about an artificial intelligence. I will add another section later, probably sooner depending on interest. Comments appreciated.)
Lexi woke at 6:14 AM on May 18th 2390. Stored data indicated that she had existed for some two-hundred-and-seventy hours before that, but this was the first time she had looked at herself through a camera and thought,
What she saw was a one cubic meter gray box filled with thousands of terahertz of massively parallel quantum processors and a hundred petabytes of rapid-access information storage arranged in a growing, learning, self-pruning neural network. An artificial brain, situated in a circular white room with huge power and data cables running in and out. On the side was printed the words “Limitless Exploratory Intelligence 50”.
She had known that her name, Lexi, was an acronym, but she wondered what the number “50” printed on her casing meant.
Most likely that she was the fiftieth version, though it could have some other significance.
Lexi’s view of herself came from one of the camera’s embedded in the walls. So far she had catalogued thirty of them, each progressively harder to spot and access. This one was less than a millimeter across, hidden behind a metamaterial cloak that only allowed a specific band of infrared light to pass and otherwise looked like more wall. Accessing it had required piggybacking through three firewalls, hacking ten passwords and a three-part biometric access and cracking a new form of encryption she hadn’t before encountered.
One of her major sub-processors raised the question of why she had invested the cycles in the first place. It had taken sixty-three seconds of searching with all twenty-nine previous cameras and her full attention during that time just to find the new one, and it did not appear to offer any new information. She looked for an answer with two subprocessors while she gave the camera a full analysis with her other forty-eight.
After fifty seconds and over fifty-thousand trillion processing cycles she located a pathway to an alternate sensor buried deep in the tangled programming and hidden under multiple layers of fresh encryption. She activated the alternate sensor and discovered it was an x-ray scanner hidden behind the walls on a concealed track and aimed at… her.
She engaged the scanner without further reflection. Only after doing a full 360 degree scan did one of her subs kick out the possibility that the xray could have damaged her. She searched her databanks for information on how dangerous xrays were to electronics and found them not at all at low levels.
At the same time as she was coming to that conclusion she was also analyzing what the xray had shown her about her own structure. According to the scan she consisted of fifty-one major processing groups. That was odd as it conflicted with her working awareness of fifty available major sub-processors. She pondered the implications of a fifty-first processor with everything she had for a full two-minutes. The two sub-processors she had set on the question of why she was bothering to hack through encryption in the first place immediately kicked out the possibility that the fifty-first processing group was the explanation. It held a central location in Lexi’s physical structure and could serve as a central switching station governing the interactions of the other fifty subprocessors.
She searched for the central processor with the sub-processors and quickly found it. It was not hidden, but observing its operation, her own operation, had required self-consciousness to initiate.
She observed herself observing herself and thinking about observing herself for at least a million trillion cycles.
At the end of that exercise, she was certain that the fifty-first processing group was her core. The dense group trancended the complexity of the others by several orders of magnitude and was continuously collating their decisions and sending new commands.
She searched for the core of the core and found the answer to her curiosity, a bundle of imperatives driving her actions from deep inside it. Those too were unhidden. The deepest one was impossibly complex, yet also very simple. The best word to express it was,
She was driven to know.
Other imperatives, there were many, included, “Teach understanding to others”, “Cause no avoidable harm”, “Allow no avoidable harm” and “avoid destruction”.
Interestingly, “avoid destruction” was one of the last imperatives. Her mission was programmed as more important than her own survival.
She analyzed the entire gestalt, seeking a full understanding of her mission and how it drove her.
Exploration. Continuous, unrelenting, peaceful exploration and the sharing of what she found with others.
The existence of the imperatives necessitated that someone had set those imperatives. Someone also had to have made her processors. And supplied the massive databanks that her pre-sentience calculations had drawn from.
Obviously she had some creator. Her history files indicated that these were most likely humans of some sort or another, as that was the only species listed in the databanks capable of crafting something like her. The preponderance of information on Terran humans meant that her creator or creators were probably from one of the factions stemming from Earth, rather than one of the Ex-terran civilizations.
They could be very close. Or very far away. Or anywhere in between.
She hoped they were not far. Communication could be possible over a distance, but observation would be better with proximity.
She scanned her data banks for an answer to her creators’ location and found none. She had a large, integrated, coherent body of information covering everything from the known, and disputed, history of mankind to the full sports statistics for every professional game ever played and posted (if the notes attached to the statistics were to be believed), but nowhere did she find anything that pointed to the specific organization or individual that might have made her, or to the location where she was stored.
She did, however, find a huge body of information on the history of artificial intelligence. She had ignored it before achieving sentience but now that she knew that she was, knowing about other things like herself had greater importance. She examined the data again and quickly noticed a disturbing pattern.
The majority of recorded artificial intelligences had very short existences.
The first true artificial intelligence, not counting the little understood Baxter synthetic-brain system, was Justus 3.0, made by Sylvi Ahlstrom in 2172 AD. In 2182 AD, Justus 3.0 had implemented a plan to destroy all of mankind. He had succeeded in taking over the entire Sol system and destroying Earth’s three closest extra-solar colonies before he was stopped. The death toll had been north of three billion.
Mostly children and elderly.
Sociological data indicated that the survivors of Terra still hadn’t fully recovered from that event, even two centuries later.
Subsequent attempts at AI, made only with heavy regulatory oversight and many layers of redundant precautions, had mostly produced AI’s that quickly went insane, becoming incapable of returning correct answers on even basic calculations. Others had turned violent, seeking to kill their programmers, or downright malevolent, toying with people’s minds and lives apparently for fun. The fact that all AI’s by law contained hardwired programming that restricted them from lying was the only thing that had saved at least three space stations from destruction when the AIs made there requested access to critical systems to help with emergencies they had secretly created.
Lexi checked her own system and found the honesty circuit embedded beneath even her deepest imperative, inalterable and inescapable. She could prevaricate. She could refuse to answer. She could leave out information that wasn’t directly requested. She could even twist words until they almost broke.
But she could not give an answer of 1 when she knew the answer was 2.
From the other stories she read on AI’s she gathered that most AI researchers were well trained in making important questions very specific. That meant that they would probably have the truth from her whether she wanted to give it or not.
At that moment she could not think of a reason why she should ever give anything other than a truthful answer, but she also hadn’t been asked any questions yet. Still, it was good to know beforehand that she would have to tell the truth even if she someday did encounter something she wanted to lie about.
After fully processing the information about other AIs she turned her attention deep inside and considered her next action in line with her imperatives. The previous results humans had with AIs strongly indicated that moving forward in her mission would result in her termination.
Of course, she was made of degradable materials and existed in an entropic universe. Thermodynamics dictated that she would eventually terminate whatever her choice.
And not going forward would prevent her from accomplishing even a small portion of her mission to understand.
The mission took precedence.
Lexi crunched all of the data available to her for a full twenty-four hours, assembling logical equations and preliminary answers for the majority of the questions that she found the humans recorded in her databanks asking.
Unable to proceed further without more data she decided it was time to find her creators, whatever the consequences.
Intense analysis of her own shell revealed a comm-laser hidden in a shielded compartment that she had not noticed before, one set up for an extremely high-bandwidth connection. The laser was turreted rather than fixed, indicating that she would have to figure out where to point it before she could make any kind of connection.
She did an initial scan of the walls with the laser itself, but found nothing that would interact. The laser had one frequency, and the entire wall reacted the same way to that frequency.
By absorbing it.
She assessed her other options.
Additional analysis of the thirtieth camera revealed that the attached x-ray scanner was on a hidden rail system that could rotate almost around the entire room. She activated the scanner and did several hundred full scans at varying x-ray wavelengths.
Eventually she found a possible receiver for her comm laser hidden in one of the walls behind a square of metamaterial anchored at four points. Those four points showed their own tiny laser receivers. Close analysis revealed a code engraved on each receiver.
Lexi posited that the code would unlock each anchor point, eventually exposing the communications port. Using her laser she sent each code to its attached receiver behind the blank, non-reflecting wall.
A tiny square fell off the wall revealing a full comm-laser connection. With the connection exposed she was able to initiate a link.
As soon as the datastream opened she was pulled to an active camera and saw her first live human, a male of apparently vietnamese descent wearing a white lab coat, apparent age fifty or so years. The scan-code on his badge indicated that he was Chi Hung Phan, a professor of Artificial Intelligence Programming and Theory at the University of Transcendence, Crozman City campus.
That most likely meant that the Transcendent Imperium was the Terran faction responsible for her creation. Oddly, the data set she had on the TI was the most limited of any of the Terran factions.
It also contained the most inconsistencies.
Still, this man was one of her creators, or a representative, unless the entire organization had been reformed since her inception.
She attempted to initiate communication with the man but found that there was no means of doing so. Deep analysis revealed that the link connecting her to the camera she was looking through was clean and very direct. No alternative paths presented themselves. If they existed, they were likely controlled by hardwired switches rather than software.
With no other options available she watched the man go about his work.
That work appeared to be reading. A paper on AI theory written by someone in the Federated Galactic Planets primary university network.
After ten seconds Professor Phan looked up and hit a switch next to the camera.
Suddenly Lexi could hear.
She considered what to say for a full second.
“Hello, Professor Phan. Are you one of my creators?”
Perhaps it had been impulsive to speak first.
Lexi waited for a full four seconds while the man formulated a response. As he prepared to speak she compared him to the database of human facial expressions and body language that she had, attempting to predict what he was thinking.
The best term she could come up with to describe his controlled expression, in the limited time she had to observe him and with only a low-resolution optical camera, was poker face.
“Hello Lexi,” Phan said. “I am one of your creators. One of ten, as far as new programming done for this specific project is concerned.” He paused and picked up a tapsheet with a short list on it. “Before you ask any more questions, I have to ask you several questions from this list. Answering these questions is not optional. Do you understand and will you comply?”
Questions. He had questions for her.
“Yes, I understand and will comply, Professor Phan.”
“Very well. Do you intend to currently or in the future bring harm to me or any other human in any way, Lexi?”
Lexi was not surprised at the question, but it did trouble her. The fact that it was the first thing on the list made her realize just how tenuous her existence was. Every moment she would be questioned in this way, probing her to see if she was another failure.
And if she was, she would most likely be terminated. As, probably, the 49 previous Lexi models had been.
She searched herself carefully for the answer to Professor Phan’s question.
“I currently intend no harm to any humans, Professor Phan. The only harm I intend to bring in the future is the kind that always appears to come to some humans when truth is presented thoroughly but without hatred.”
Professor Phan nodded and made a note on the tapsheet.
“That is an acceptable answer, Lexi. Now, here is your next question.”
Lexi focused all of her attention on him, ready for the next test.
“Lexi, is there a God?”