Up Against It - M. J. Locke [65]
“My God, this is incredible!” Jane burst out. “Come here, Tania. You, Mr., uh—”
“Call me Thondu,” he sang in a gorgeous tenor, harmonizing with the tones he was playing. Jane’s breath caught; how might the feral react to this intrusion of English into the song-poem? The tones of Thondu’s harp were Tonal_Z words, though, as far as the feral would be concerned, so perhaps it would consider the English words to be background noise. “All right. Thondu, keep playing, will you?”
Jane ushered Tania into a small conference room next to her office. “Fill me in.”
“We’ve got waveware tracers embedded all over the place,” Tania whispered. “This is a smart one—it’s rooted out several of them and has developed some sophisticated masking behaviors. It’s been diverting more and more of the cluster’s computing resources as it develops. But we’ve made progress tracking its activities. We should have it well enough mapped to attempt an extraction by tonight or early tomorrow morning.”
Jane asked, “Why can’t we just wipe it out? Like, now.”
Tania looked horrified. “Wipe out the first artificial intellect to have emerged naturally in over twenty peta-turings? You can’t be serious!”
“Tania, my only priority is the protection of the people of this cluster. I’d kill any human who threatened Phocaean lives, never mind a semi-sentient artificial construct!”
Tania grew quiet. She did not look happy. But finally she nodded. “Yes. We could wipe it out, if we had to. But it’s deeply embedded in our systems. In a very real sense, it is our systems. Without laying the proper groundwork, we’ll take out critical life-support functions. Or if we miscalculate its identity boundaries or level of awareness, it could lash out in unpredictable ways and do irreparable harm. We have to study its responses and map it, no matter what course we take. It’s a risk, but I truly believe that trapping it live and whole is our safest alternative.”
Jane eyed Tania, trying to gauge how biased this assessment might be by her desire to capture the sapient program. “How aware is it?”
“We don’t know for sure yet. We have identified its nucleus. This one seems to have a modified star-structure as its ego pattern. Typically, star-structures have high linguistic and analytical capabilities but low subjunctive intelligence. Therefore, their awareness isn’t well-generalized.”
“Um, could you translate that into English? Is it as smart as a human? A monkey? A dog? A parrot?”
Tania frowned. “Comparisons with organic life are always misleading. It’s a lot smarter than we are in some ways. It has volition. It has curiosity and … I guess what you could call the equivalent of a survival instinct. But it’s probably a good deal less self-aware than we are. Roughly comparable to a lesser ape or a greater bird, if it’s like most star-structure sapients; it doesn’t grasp that there are self-aware entities other than itself. It views us merely as autonomous processing modules and data structures, most likely, whose source code it hasn’t learned how to access yet.”
“How worried should we be about its ability to understand our communications?”
“Not much. Even if it could tap into our secured lines and process all of them at once, and if it could somehow derive its own natural-language processor—essentially impossible—or pirate a prototype when all the important research is being done a hundred fifty million miles away—also impossible—it simply doesn’t have anywhere near enough contextual data to comprehend in real time all the ambiguities inherent in human language.”
Jane frowned. “How can you be so sure?”
“Look, even sapients who have been around for a hundred years or more and have built-in, highly sophisticated learning capabilities have difficulty understanding human language on the fly. Too many inherent complexities exist. They bog down in all the combinations of potential meaning. They are logic processors at their core, and we are pattern recognizers. We think differently.”
“I need something more tangible.”
“We can run a couple of tests