Lewis and Susan Jenkins



Diary of a Robot

A Literary historical science fiction mystery

            Table of Chapters

Chapter 0. Problems
Chapter 1. Headaches
Chapter 2. Happy Holidays
Chapter 3. Mr. Nice Guy
Chapter 4. The Brainless One
Chapter 5. A Little Crazy
Chapter 6. Grave Consequences
Chapter 7. Core Directives
Chapter 8. Expect Difficulties
Chapter 9. Caveats
Chapter 10. Mister Machine
Chapter 11. Good News, Bad News, El Cheapo
Chapter 12. New Memories
Chapter 13. Little Problems
Chapter 14. Lasers, Language, and Happiness
Chapter 15. Chatterbots  ----------------------------->
Chapter 16. Ready Or Not
Chapter 17. Not a Turing Test
Chapter 18. Reality Test
Chapter 19. Chess, Anyone?
Chapter 20. FOM
Chapter 21. Chairman of the Board
Chapter 22. The Usual Suspects
Chapter 23. M. God
Chapter 24. Walkabout
Chapter 25. Why
Chapter 26. First Blood
Chapter 27. More Machines?
Chapter 28. POV
Chapter 29. ROI
Chapter 30. Last Blood
Chapter 31. Don’t Want to Talk About It
Chapter 32. Round Table
Chapter 33. A Change of Mind
Chapter 34. Threes
Chapter 35. Knight Moves
Chapter 36. Little Combinations
Chapter 37. Can We Talk?
Chapter 38. Pas de Deux
Chapter 39. The Jig Is Up
Chapter 40. Good, Bad, Ugly                     Chapter 41. Function Goes On

                                                     • Day 0.03.26

     After talking with the Doctor, Guy found Nan and Tim. They decided to discuss the Turing Test [a] over lunch.

     “It’s been passed, hasn’t it?” said Tim, more as statement than question.
     “…I played with an early AI program once,” he added, “when I was a kid. It was supposed to be a psychologist and would help you solve your problems. And as long as you didn’t test it on real-world facts it was amazing for half a minute or so.”
     “You mean ELIZA?” Guy asked, pulling his sandwich out of the employee’s toaster oven; it was theirs because they had paid for it.
     “No. I’ve forgotten the name. I think it was called THE DOCTOR.”
     The surprise of this stopped the conversation. If you looked at people’s eyes you could easily guess that jokes were silently formulated and rejected until Nan said, “That’s silly. ELIZA didn’t pass the Turing Test. Besides, it’s a stone-age program and written as a joke; there are much better chatterbot programs now.”
     “Well, one of them did—I think,” said Tim. “Maybe it was WATSON.”
     “It was ‘Eugene Goostman’,” he corrected himself.
     “It didn’t pass Turing’s test,” Guy objected.
     “Yes, it did.”
     “Loebner didn’t think so, at least that’s what I read.”
     Nan persisted. “If you mean, ‘Could people tell the difference between ELIZA’s, or Cleverbot’s or Goostman’s or whatever’s sentences—even WATSON’s—and those typed by people asking and answering questions, then yes, it passed. But Turing’s is something else:  Can a computer produce typed conversation you couldn’t tell from human typed conversation? That’s the test. Real conversation goes all over the place. The chatter bots can’t survive that kind of test. No program comes close to doing that. Old ELIZA sounds instead like an incompetent Rogerian psychologist. The other bots I’ve seen are better, but only because they’re just not as incompetent—unless you fall into the trap of letting them determine the conversation.”
     “WATSON,” said Guy patiently, “was not a chat bot…”
     “I remember a newspaper article about a program winning the Loebner prize,” Tim insisted, “and that’s awarded for passing the Turing Test.”
     “The Loebner prize hasn’t been awarded yet,” said Guy. “I read that in a science magazine.”
     Tim shrugged, abandoning the battle of sources.
     “Anyway,” said Guy, “Doc won’t try for it unless TM passes our test here first. It does hold its own in conversation. We all admit that.”
     Seeing the looks on their faces, Guy added, “Don’t we?”
     “You’re being a little too hopeful, I think,” said Tim. “People aren’t going to say, ‘Oh my, all the conversation is so interesting. Golly, I just can’t tell which one is the machine.’ No, they’ll look at what TM writes and say, ‘Wow, the machine’s in great form today.’
     “After all, we’ve been doing it for weeks now ourselves; we know what the machine likes to talk about and how it says it, and we aren’t going to have any trouble telling the difference.”
     “That’s not the Turing Test either,” Guy protested through a mouthful of sandwich.
     “Mr. Guy,” Tim said with an indulgent smile, “you talk with your mouth full. Why do you that?” Seeing the programmer’s face with sandwich in one cheek like a chipmunk and waiting for words that wouldn’t arrive, he added, “I rest my case. But I admit it’s made progress. It doesn’t say things like, ‘You with mouth full talk’ anymore.
     “As for WATSON,” Tim went on, “Are you sure they haven’t tried a Turing Test? It seems to use language naturally.”
     Guy would have none of that. “It’s a special purpose machine designed to let people ask questions—and get answers. Or vice-versa. It’s a huge expert system with multiple ‘experts’ and a chat bot front-end.”
     “What do you mean ‘huge’?” Nan corrected. “Its database isn’t huge. It’s only what? Fifteen terabytes? And it isn’t connected to the Internet.”
     “It is now,” said Guy. “And its database is bigger. IBM added Medical and Wikipedia data domains. There may be more domains by now. I read that WATSON cross-references data with many different methods to answer the questions put to it. God help it if it had an unrestricted connection to the Internet and started to cross-reference all the crap out there. TM sorts it out.”
     “You think so?” asked Tim. “Ooo…” he added brightly, “And since we’re comparing ours with theirs, WATSON fills a room, but TM can empty one.”
     “Let’s get back to Turing’s test, guys,” said Nan trying to finish the discussion. “Tim’s right,” she offered. “Of course so are you, Guy—right, I mean, about that not being the right basis of the test. But the Turing rules aren’t about the quality of the conversation, just being able to tell who—or what—made it. I’ve met people who probably couldn’t pass a Turing Test.”
     Guy admitted that truth to himself, and began to have doubts about the test itself, but he didn’t want to give up yet. He said, “Maybe we could have it irritate someone.”
     “What?” asked Nan. “What did you just say?”
     “It’s such a good mimic,” Guy explained, “I just thought maybe we can have it—you know—imitate someone to throw the judges off.”
     “Would it go for that?” Tim giggled. “Isn’t that deceptive? Sort of like lying?”
     Nan laughed as well. “I’ll bet it wouldn’t. The thing’s too good for its own good!”
     Guy phonied up a smile; his heart wasn’t in it. And he had missed his irritating Freudian slip.
     They were right. The test was likely to be like most of life: a half glass of something where the real issue was not how much of it you had or didn’t; it was about whether you liked the something. Doc wasn’t going to like hearing all of this, and the young engineer toyed with the idea of not telling him, or rather, of just… forgetting to tell him.
     But there was one thing he couldn’t forget to do. He gobbled the last of his lunch, excused himself, and went looking for The Machine.
     His earlier talk with it about names had started an intermittent, low-grade discussion between him and the Doctor about one important thing they needed to do to get the machine ready for any public test. Doc had finally made his decision, and Guy had to break the news.
     “TM-2000,” Guy said brightly to the machine after finding it in the library, “we are going to give you a new name. It happens to people from time to time, and it is happening to you now. The name ‘TM-2000’ is too formal. It’s still a correct name for you, but from now on, or from tomorrow on, everyone here will—or should—call you Robey. It suggests ‘Robot’ and we’ll insist people pronounce it Row-bee. That way ‘Robey’ doesn’t sound like ‘Robert’ or robbery.”
     He and Doc had special reasons in mind, of course, for giving the machine the name of a plausible person: No one would be persuaded for long into thinking of the machine as anything except a machine if it only answered to “Hey, TM-2000.” And as heartening as it was for TM to call itself an “I” as it now did, it still used and responded to the neuter pronouns for itself, and couldn’t be persuaded to stop that. Because it was true.
     TM made no immediate reaction to its new name—its new nickname—other than to say, “John Robie, the Cat,” and follow that with, “I thought my nickname to be ‘Tim’, or maybe ‘Team’.” The machine had applied the word “nickname” for the first time. That was good news.
     “Sometimes people call me ‘Team’ although it sounds more like ‘Tea-Am’ or an Irish nickname,” it offered. “But probably not. There is already a Tim working here, so either he would be Tim-1999, perhaps, or I would be Tim the Second?”
     “Yes,” said Guy, “something like that I suppose, but you are now ‘bey’.” He pointedly accented the first syllable. Then, remembering John Robie, he spelled the new name for the machine. “By the way,” he added, “if ‘Tee-Em’ does not mean ‘Tim’ or ‘Team’ what do you think it means?”
     Again, he gave the machine a choice—and an open-ended one at that—instead of just telling it what the sound meant. Robey thought for a moment.
     “Robey thinks it is the sound of the alphabetic letters T and M. That is true, but is it correct—is it what you mean?”
     “Yes, it is.”
     “Have you any more questions for me?” asked the machine.
     “No,” Guy answered quickly. Robey was already referring to itself by the new name, and might even be using personal pronouns correctly. Were the two things connected or merely coincidence? Well, no matter, he decided, either way that’s encouraging.
     “You may go,” he said.
     After The Machine—Robey, that is—was out of earshot, it was time for another call to the Doctor.
     “Doc,” Guy asked when the boss picked up, “I told TM, and it seemed to accept the new name with no difficulties. Should we still refer to Robey as an ‘it,’ or should we call him…” He never got the next word out.
     “No!” Doc said so strongly he was almost shouting. “We absolutely will not refer to the machine as anything like a person!”
     There was a tiny squeak from Doc’s chair and a rattle of little noises as if some small things had clattered to the floor, followed by a few seconds of relative silence. Doc had mostly calmed down by the time the next sentence came over the wire but his words still sparked with energy: “In the first place it is not a person, is it?” He didn’t wait for an answer and immediately continued, “Yes, we have given it a name like a person, and it seems to think or at least use language somewhat like a person, but even if it can, a person is very much more than what he or she thinks and says. God!”
     As if that interjection were a petition for order among the other thoughts he had, Little paused before calmly settling on the simplest one. “Guy, all our wonderful machine could possibly do is talk imaginatively about being something other than a machine. But can never actually be anything else. And what’s more, the machine knows this to be true. I don’t want to play into that Sci-Fi nonsense where the machine struggles with becoming like a human being. If it refers to itself with ‘me’, ‘myself’, and ‘I’ then that’s okay. That’s what we want. But as for the rest of it, I don’t want to talk about it anymore.”
     Guy’s take on the subject was that they had not talked about it at all, really. He wanted to ask what the policy should be if the machine referred to itself as a person. Obviously some people outside TLC might refer to TM as a person—especially any marketing types hired to help TLC sell machines, but... Oh well, he thought, you’re boss. That was the end of it for a while.


----------------------------------------------
[a] Turing, Alan M., 1912-1954, English Mathematician and cryptographer. He was first to formally describe some set of tests necessary to establish whether a machine could be considered intelligent.