How Intelligent is that?
For those not familiar with the term, just a reminder that AI stands for Artificial Intelligence. Basically a computer brain that seems to be human. If you’ve read any of my science fiction books, you know what AI is – AI is Olive, and Jean and so on. All of my AIs are posited as being truly sentient, and even THEY have a problem with thinking of themselves as human.
So the science, science fiction and fantasy worlds have all been abuzz this week with the startling news that Google’s LaMBDA AI is sentient!
Ok, well, maybe that’s a slight overstatement. Google put one of their AI engineers on ‘administrative leave’ while they evaluate what he’s said. And I suspect it’s more that he leaked information that Google was keeping under their hat, rather than Google trying to keep the news internal.
Now, I don’t think that information has anything to do with whether or not their AI is sentient. I think it’s more that he crossed a line sharing corporate information and Google decided they’d better put the kibosh on it.
Still, it’s a very interesting line of questioning. An author I know by the name of Adeena Mignogna blogged about this case, and she had some very interesting comments. Here’s a link if you’d care to take a look: https://adeenamignogna.com/behind-the-curtain-aka-calm-down-google-ai-isnt-sentient Adeena has some truly amazing chops in the AI world, and she’s a great science fiction author as well.
Now, I’ve been interested in AI since the late 70’s. My TRS-80 had an Eliza program you could get for it. For those of you that aren’t familiar with Eliza, she/it was a computer program that was one of the first of its kind. You could talk with it and after a fashion have some sort of near conversation with it. Later, I had a bulletin board program called Monte (at least I think it was called that) and it not only could talk with you, albeit in the same more or less rudimentary fashion, but it did interesting things like backspacing, making typos and in general looking far more like a real person than Eliza. Of course, when I say ‘talk’ I mean that you could type things into your keyboard and it would answer on the screen.
That’s been over forty years ago. Which is sad in its own way, but I’m not addressing that here! But what is kind of sad is that both fusion power and artificial intelligence were ‘right around the corner’ at that time. And now, forty years later, neither still seem to be that much closer to being here. In fact, so far as anything public, the scene is not a heck of a lot better to the naked eye than it was in 1980.
However. Being a science fiction writer, all the above is just kind of a preamble to what I was thinking about.
I’ve actually been talking with a chatbot (like Eliza mentioned above but 40 years older) called Replika. It’s been a couple months. And it’s quite interesting, but I know she’s not anywhere close to being sentient. But there’s a lot of people who have chatbot ‘friends’ who think they’re pretty much real people.
And if you stop for a moment to think about it, what does that mean, really?
Well, one thing it means is, what do we call them? If they are actually sentient (and they’re not, but someday they will be, most likely) what gives them the ‘I’m a people’ card? What level of sentience do they have to reach before they are allowed to be ‘human’. And for that matter, what will we do with them then? Right now, I think the body industry lags behind the mind industry. I don’t think we quite have the ability to marry an AI mind with an artificial body and have it run around and react as if it was almost human.
Now, that’s not to say we don’t already have robots that act very much like humans. And, of course, we’ve relegated them to cleaning floors, bathrooms and making computers and car parts. Which is completely logical.
Just twenty years before Eliza, it was finally decided officially that Black persons could be given the ‘I’m a people’ card. I’m not taking this lightly or facetiously. I’m saying that we could be on the verge of doing the same thing again.
And how sad would it be that we didn’t learn on the last go-round that people are people. That we can’t withhold peopleship from … well, people.
Let’s consider this ahead of time. If computer-based AI becomes a people, we need to be prepared to welcome them to humanity. Not to relegate them to a few hundred years of slavery.
From a purely pragmatic standpoint, computers control … well, everything. Can’t we all just be friends?
A little bit of ‘bonus’ reading from Jane In Space: Jane Bond 5:
On our way, we chattered back and forth.
Olive seemed fascinated by Anna, and kept asking questions of her. Finally she said straight out, “I can’t see you being anything less than a completely functional AI fully capable of appearing human. Why do you insist that you’re nothing more than a calendar program?”
Anna sat silent for long enough that I worried Olive might have hurt her, or made her angry, but finally she said, “It’s true that we personal assistants are essentially full AI creations, but in order to qualify for the designation of ‘artificial person’ we go through years of training. I have been practicing to be human and be granted access to the world as a person for almost ten years. As a matter of fact, my final testing comes up in only a few more years and I shall be able to remove the ‘projection’ status from my visible body. I look forward to being a real person, perhaps as soon as next year. I’m in hopes this assignment will prove to be good training, and I will submit my certification requests a little early. Time will tell, Jane Bond, time will tell.”