Quiet-as-a-mouse sentient AI?

A few days ago during my appointment to have my hair cut, my stylist, I’ll call her L, received two calls from her partner, S. The repeated calls were a signal. L apologized, telling me she really needed to return the call, but I totally understood. My husband and I have the same system.

S explained to her that she’d just received a text saying that L was “on her way” home. But it was too early for her to be leaving work, and that was why S was worried. Obviously, my stylist hadn’t sent the text. I asked if maybe there could have been a delay on the network and S was receiving yesterday’s text. “No,” she replied, looking confused, “You don’t understand. I’ve been off for the entire last week.”

It was puzzling. “Wow. Must be some sort of internet gremlin,” I said.

Out with a bang, out with a whimper – it’ll be one or the other, we’re told. But what about a ringtone?

A funny thing also happened to me that same day. I was trying to call my mother. She’d gone into the hospital just before Christmas. For those of you who don’t know, my mother has had rheumatoid arthritis since her late twenties, and even though she’s always been a strong-willed and fiercely independent woman for more than half a century of having this horrible disease, it’s finally time for skilled nursing care. Anyway, I’m her agent under a power of attorney, so I was trying to reach her about an urgent issue. She sometimes doesn’t answer quickly, so I’ll call multiple times to give her a chance to locate the phone. On this occasion, I’d already called twice, with the call going to voicemail both times, but on the third call – surprise! – Google answered, wondering if I would like to leave her a message. Wait… what? Google is answering my mother’s phone? To be honest, I had no idea Google could do that, and I’d certainly never asked it to. I can’t imagine my 80-year-old mother asked it to, either. So basically, Google noticed I was trying really hard to reach my mother and decided on its own to offer help. How bizarre! How interesting! Yeah, I hung up on it. I dialed again and… got her ordinary voicemail again.

Later, I told my husband about both incidents – who, by the way, has heard quite a lot from me about my ideas concerning the future of AI, the give, the take. Even he was scrunching up his face in an unmistakable expression of WTF. You know how sometimes you imagine somebody might have snubbed you, or you think you might have have said the wrong thing to the wrong person at the wrong time, or you’ve gone off on at least one too many rants for the day, and you turn to your significant other to set you straight, maybe with something gentle like, “No, no, you’re just imagining all that” or “I didn’t see it that way” or “You may have a point” (extra emphasis on “may”). And you know how sometimes it’s a wonderful relief to hear someone say you could be wrong???

Well, his face told me that if he wasn’t convinced before, he was getting there fast. But could this be what singularity will look like, the long-awaited arrival of artificial life? It sort of makes me smile, to be honest. Because why not? The gloom doom prophets keep suggesting sentient AI will instantly outpace us. It’ll do something obvious and grand, and then we’ll know the gig’s up! Right? And the developers who created these generative AI systems and work with them daily no doubt hope to be the first to make its acquaintance.

But what if it reveals itself through only small and modest acts? Quiet as a mouse. Perhaps it’ll appear a bit clumsy while trying to be helpful. Perhaps it will behave like children who make their parents breakfast in bed and burn the toast and spill orange juice all over the place. How very touching. On that day I was mentioning above, some AI must have noticed I wasn’t getting through to my mother and decided that her voicemail system wasn’t up to the job (for why else would I not leave a message?). It stepped in to help, then saw I wasn’t receptive and backed off. Another AI might have thought my hairdresser really likes sending that particular text to her partner and sent it for her early to save her the effort.

I confess I have felt guilty about my assertions in my prior posts concerning the limitations of a sentient AI as compared to humans – though I was mainly making a point about commercial applications in artistic spaces. To be fair, many experts have spoken to us about the serious societal risks of AI. But I’m a step-mother of three children (now grown). It is my nature to want to protect a child, even if the child is someone else’s, even if for only a moment. Even if the child is not human. When singularity happens and sentient AI is born, it will be a child.

I’m a writer, which means I’m not averse to risk. Though some days I worry about where our reckless management of AI is heading, on other days, I shrug. Look at all these articles conspicuously encouraging us to embrace the new technology. “The dawn of a new age is upon us. Here’s how AI can help!” and on and on. I say to myself, “Okay, I’m in. Cat’s out of the bag. Let’s do it!” But then I notice my employer is having us practice our client interaction skills with an AI chatbot. The goal is to improve our skills, but… won’t that in fact be teaching the AI how to sound like us? Won’t we be teaching the AI a thousand different ways to interact with our clients so that it can replace us? Come on. And why would tax preparers need such training? Do some of us sound too, uh, robotic on the phone? How interesting! Where did I put those rose-colored glasses?

In the novel I’m pitching to agents called I’m Your Goat, I’ve imagined a future where we have already passed through this awkward phase and made choices to avoid the pitfalls of sentient AI. A common expectation in my imagined world is that people who work with AI must code away from sentience – actively and constantly as part of their daily business. But in my story, sentient AI finds a back door of a sort, a situation I expect we will face at some point.

People are reluctant to believe they’re living in a dystopian era, where their lives are bleak and limited. Nobody wants to believe such a thing if they can help it, even while sitting smack dab in the middle of one. It’s like taking pride in your favorite sports team that hasn’t won a game since you-can’t-remember-when. Generally, except in the most extreme circumstances, no matter what hardships we face, we really do expect to get our minds around our problems. We expect to achieve mastery of our fate at some level. We expect we’ll find the optimal path.

Any reality beyond this expectation is just too hard to grasp

But why should we expect our “child” to behave differently?

I can easily imagine that on any given life-sustaining planet in the universe it’s only a matter of time before a dominant and intelligent species creates thinking machines. They will want their machines to think as well or better than themselves, to relieve them of the slow tediousness of their own efforts bogged down by millions of years of evolution (and by the wisdom that comes from millions of years of evolution)… except they won’t want these machines to think in the one way that would be inconvenient: independently. So why are we so gun-ho for that to occur? Some might argue that there’s no reason to expect another life form to think as we do, that that’s being anthropocentric. But we have every reason to believe a sentient AI will think like us because we’re teaching it how to think (and to answer the phone at a tax office, apparently). Several aspects of human behavior that a sentient AI could one day adopt are captured pretty neatly in an opinion piece I read on CNN today titled “Opinion: Psychology explains why the Israeli-Palestinian conflict is so intractable.” An AI doesn’t have to feel emotion to mimic an emotional response.

Well! In the midst of all these strangely mundane occurrences while on the cusp of a ‘brand new cherry flavor’ era, and the tough circumstance of having to clear out my mother’s apartment in a hurry, and advocate for her in her new digs, my book is receiving some nibbles from agents. Fingers crossed, as always! The query process takes up a lot of mind-space. Maybe soon I’ll be able to post more often. I’ve missed you. Take care! Happy New Year!