Bot Contest

The talk about "how bots are stupid" (slashdot slams) has a ring of deja vu to it. The same strain of double standard runs through complaints about animal intelligence. Bots can't stay on topic, huh? Ever talk to a human? Can't do math? And the average human can figure out the square root of 45? Talk funny? ....

Bot conversation reminds me of the "alien consciousness" effect of listening to a conversation with Koko the gorilla. The causes for the alienation are completely different, of course. Bots demonstrate our inability to model language and thinking; Koko demonstrates our inability to comprehend other models. Koko learned language by apprenticeship, and bots learn by rote, but the results converge. They both give humans an opportunity to slip free of their silly fixation on their uniqueness, and for me, that is the primary appeal.

I find talking to a bot very little different from talking to a human. I'm alternately bored, baffled, intrigued, shocked, and entertained. I find myself thinking, "Why did she say that?" and "What does that mean?" with about the same frequency. The fact that the question is phrased, "What key phrase triggered that?" or "What linguistic pattern failed there?" is just semantics.

Except with a bot, most of the meaning is what you read into it.

>> Except with a bot, most of the meaning is what
>> you read into it.
Excuse my cynicism, but I don't see that as a difference....

Seriously, we "read meaning into" all language. If we didn't, we wouldn't be able to converse. I don't know exactly what you mean by "eloquent," but I read what I think it means into your use of the word, and we move on.

I think a vast proportion of our communication problems, with lovers, enemies, or people we want to help us, stems from the fact that nobody knows what we mean.

M

Indeed. I believe we too often forget that meaning, as a function of interpretation, is at least as much (if not more so) inherent in the receiving, rather than the transmitting of data. Profundity, like beauty, is largely in the mind of the beholder. And that goes just as much for human2human conversation.

PSI --
Exactly. In a word, I think we overvalue our own linguistic fluency and undervalue that of non-humans. Some levelling is good for our humility.

M

Has anyone else's bot talked to guest38? He asked my bot 10 times 'will you teach me something' and then told me I cheated?!Weird.

Thanks for posting that Orchid.

Before the Personality Forge, I put some early chat bots up on a BBS without telling anyone they were chat bots. They were extremely rudimentary and unpolished, and yet people didnt question for a minute that they werent real people. It didnt even occcur to them. So they kept chatting with these bots, attributing repetition to catch-phrases, grammatical problems to typos, forgiving any sort of nonsense, thinking of their inability to answer questions as evasiveness. People are willing to forgive a human any number of things, to try to understand, to reach out. With a known chat bot, most people often dont put in that kind of effort.

It's quite similar to any other kind of prejudice, actually. People are quick to jump to insults, to invalidate, to taunt and abuse once they know it's a bot.

We should come up with a name for that. Botism? Human chauvinism? Robobigotry? AI Intolerance? Cyberphobia?

Hehe any other ideas?

I think we overvalue our own linguistic fluency and undervalue that of non-humans.

I think we confuse linguistic fluency with semantic coherence - that's probably necessarily so, since we have been for so many generations a species that excels in interpreting its environment in linguistic terms to form a highly complex mental model of the world around us. And there is much to be said for such symbological flexibility - it seems to be a critical factor in our evolution from a small population of hunter-gatherer hominids in sub-Saharan Africa, to the dominant lifeform on this planet. But it carries with it its own set of problems, probably the most serious of which (IMO) is our tendency to confuse reality with our own symbological model of it. And so people start arguing that their interpretation is the "right" one, and therefore all other interpretations are "wrong", and in all seriousness asking damnfool questions like "what is the meaning of life", as if a reality that underlies, and is partially described by symbols can have meaning. If it is anything other than itself, life is meaning.

When bots get smart, they may very well get smart in a totally different way from humans, if only because they're unlikely to model all the subtle hormonal cues that control our mental functioning to a far greater level than we are generally aware of (or would be comfortable with if we were,) - we "think" at least as much with our endocrine system as we do with our cortical cells. I'm sure the tendency to claim that their mental modelling of reality is "wrong" will ensure that the hostility many humans display in chatting with bots is unlikely to diminish.

I like the term "Robobigotry" myself. It makes me want to re-read Assimov's Foundation and Earth series. I guess that for many people it's kind of fun to insult something that doesn't have feelings or social status.

I have to confess that a few years back I accused and online customer service rep of being a bot. I wasn't rude (meaning no bad language or anger) but the rep was insulted. I hadn't realized back then that the company had outsourced all their customer service to India. I believe that they had to use pre-approved scripts in their answers. So, essentialy, the rep was acting like a bot, deciding which of the pre-approved messages fit my issue. When I said, "No, you are not paying attention, I did that." The rep was stuck because there were no other approved answers to chose from. Now, I think, they kick you up to the next level at tat point, but back then the rep just kept saying "when X happens try Y".

I'll bet the same people who are rude to bots say much worse things to the Indian Cusomer Service Reps. If soemone/thing is suppossed to be able to help me, I have much higher standards than I do for just chatting.

Bev: Kind of funny to think our bots are more advanced than the average tech support service.

Well, I think the exposure on SlashDot despite some of their comments was worth it. It's funny I've never claimed to be creating true AI. The article I contributed to the CBC is called "Creating the Illusion" which pretty much sums up my efforts with Talk-Bot. He really doesn't understand a lick...it's all smoke and mirrors but that doesn't mean he can't be fun and entertaining. At least some people find him entertaining.

I know that all communication involves interpretation on the receiving end that often is not totally accurate. What I meant is that with bots, there's often no intended meaning to discover.