webfarmer | Reading

Some thoughts about digital personalities in my work

(self-indulgent blather about my take on artificial digital beings, as I’ve written them)

I’ve been watching the latest AI developments with a somewhat…oh, what word do I want? Not jaded, not cynical, but definitely somewhere in between. Especially when I start reading about “AI Grandmas” and the use of that tech to speak to long-dead relatives. Oh, it’s presented with that same amber glossiness that seems to dominate the worlds of AI visual creations. But…we’re already seeing some of the dark side of these AI creations with reports of self-harm and worse coming from AI “personalities.”

One reason for my attitude is that the creation of self-aware digital personalities is something I’ve somewhat explored in my work, most notably the Netwalk Sequence series and the Martiniere Family Multiverse Saga. In both cases, the tech I explore is already somewhat different from what we are seeing. I don’t go into the nuts and bolts of just how that self-awareness ends up happening (well, a little bit in the Martinieres). But nonetheless, I think this dynamic of what that really looks like is something very much overlooked in the current hype around “preserving the memories of your loved ones” in order to recreate them in a digital simulation. I can oh-so-easily see how it could turn bad.

What happens if AI Grandma is toxic? Or if AI Grandma develops sufficient self-awareness to start meddling in the affairs of her descendants? It’s entirely possible. And while AI Grandma might not have the ability with current tech to really muck up her descendants’ banking and financial history…there’s still a lot of damage she can do to living beings.

The Netwalk Sequence was my first exploration of just what the problems with a separate digital personality creation could end up being. I started building the Netwalk Sequence world back in the ‘90s, when digital personality uploads were somewhat the fashion in fiction and in theory.

My base assumption was that digital personalities could completely upload to the internet upon their death. In that world it’s entirely possible to be a complete personality online, with full body immersion, using the mechanism of a highly sophisticated wireless communication chip implant called Netwalk. Uploading came later, in the midst of a dramatic political struggle where an older leader—Sarah Stephens—uploaded upon her death and began to stalk and attack her opponents. The new development was called Netwalk, and the uploaded personalities called Netwalkers.

A restraint that I created in the Netwalk universe was that Netwalkers would go insane and turn predatory on living beings if completely cut off from sensory inputs. They would attack alive users of Netwalk in order to gain sensory exposures and recharge themselves—as well as fulfilling agendas and settling resentments that hadn’t been dealt with in life. In some cases this would end up as possession of the living being by the Netwalker. As a result, with the exception of a handful of rogues, Netwalkers ended up being tied to a living host, most specifically that host’s Netwalk chip. In the Sequence, we see is how this plays out within one powerful family, the creators and controllers of this technology. With some other dynamics thrown in as well—the control of a war machine of unknown origin which has some influence on the development of the original Netwalk, plus intensely weird family history that involves a lot of infighting and struggles over who controls what.

There’s no grudge like a family grudge, shall we say?

In the Martiniere Multiverse, I postulate something closer to our current concept of the “AI Grandma,” where videos and recordings lead to the creation of digital thought clones. Thought clone appearances in the Martiniere Multiverse aren’t constrained to computers and devices, however, and they can hop universes. This is somewhat connected to a magical Fae origin which is tied to a computer worm that can also skip through assorted multiverses.

The Martiniere digital thought clones (digis for short) differ from Netwalker personality uploads at death in that they are specifically digital constructs of a once-living personality, and only become activated upon specific actions by a living person who is keyed into the algorithm. The digis are fully aware that they are digital constructs and are not the uploaded personality of the dead person they’re modeled after.

Digis don’t appear in every Martiniere book. To follow their development chronologically in series order, start with The Enduring Legacy, the fourth book of the Martiniere Legacy series. We see Gabriel Martiniere’s first awareness of digis shortly before his death, when he ties the appearance of a dangerously destructive computer worm to specific holes in not just his memory but the memories of his closest family. Gabe takes the first steps to establish the bounds of his digi, with a specific activation algorithm tied to certain family members.

More details about digis and their creations happen in two of the Martiniere Legacy standalones, The Heritage of Michael Martiniere and Justine Fixes Everything: Reflections on Mortality. Heritage shows Gabe’s activation; Justine goes into further complications. However, the most details and the most explicit multiversal version appears in the three books of The Cost of Power: Return, Crucible, and Redemption.

Like Netwalkers, digis are capable of possessing living beings and bending them to their will. There are malign digis and beneficial digis. We only see them in the context of one, powerful family because, in both cases, the artificial entities serve as chess pieces in ongoing family battles. They are obstacles that need to be navigated and overcome by the protagonists.

(Sarah Stephens and Philip Martiniere would probably strongly disagree with me but—nothing says that they are pawns.)

Back in real life, Netwalk is probably not at all feasible, though digis…may be. Current technology doesn’t allow for digis to function the way I wrote them in the Martinieres, but some of the same issues raised by both Netwalk and digis still exist. The news has multiple examples of people being influenced by AI interactions to do harm, whether to themselves or others. Or of people who develop a strong emotional attachment to artificial beings to the detriment of their attachments to living beings.

Rather than the apocalyptic stuff I postulated in the Netwalk and Martiniere books, that’s the real harm in uncritical adoption of the creation of artificial beings. At what point do we slip from a clear awareness that “this is a creation; this is not real” to uncritical acceptance of these creations as real beings?

What happens if we start treating these AI creations as something above and beyond an artificial construct?

What rights will they have as opposed to living humans? Or lack of rights?

What happens if they turn malign, either due to the manner in which they are constructed or due to abusive treatment from living humans? Then what?

All food for thought.

Meanwhile, the artificial beings I created in my own worlds are definitely not your happy-happy AI Grandmas. And at times, I wonder if those imperfect visions of mine may end up reflecting an actual reality.

We shall see.

So far it's been an open winter. We've had some snow, some frozen ground, but now things have warmed up a little bit, thanks to a series of storms brought in by an atmospheric river. We don't get it nearly as bad here as the people on the West side of the Cascades will, but all the same...it should be snowing here, and it isn't.

It's somewhat of an adjustment to not be worrying about a hard-keeping horse in the winter after twenty years with a hard keeper. Even in an open winter like this, I had to keep an eye on Mocha to make sure she wasn't losing weight, and she would have required a blanket or rain sheet on some of our stormy days so that she wasn't burning calories keeping warm. But now, with Marker, it's...simply not an issue. He went from a dry lot with hay nets (to keep him from wasting the hay. Boy is a very messy eater) to the big field and...has gained a little bit of weight. He's back to the next-to-the last hole on his rear cinch, and even then it's snug. Might have to punch another hole in it to fit.

But it feels weird to not be thinking about the blanket dance.

Marker had to take a little bit of a break because he strained a fetlock (we think) in his left hind. He recovered from that quickly. But I started poking around, and thought the root of the problem might be higher up, too, so I started putting liniment on his gaskins. Which he likes but...he also has shown a taste for the liniment itself. I discovered this the other day when he tongued open the cap and was licking the bottle. Spilled some today and he was licking it off of the truck tailgate. It has juniperberry oil in it so I suppose that's what the attraction is.

In any case, we're getting back into full work, riding in the field because it has the best footing. I haven't been taking him on the roads because the type of storms we've been having reduces visibility, and even with a bright purple quartersheet I'm not sure we can always be seen. At least it's not the driving, steady, pounding rain of Western Oregon. But intermittent showers can still leave me cold and wet at the end of a ride! Especially on a windy, blustery day.

On the other hand, it's no worse and actually better than some of those stormy days I spent skiing at Timberline. I use the same type of base layers out riding that I used when skiing--synthetic, yes, with the ability to wick moisture away from my skin. In some cases, those base layers are the same ones I used in those ski days. Couple that with some of my old ski sweaters and an old hardshell parka, and it's reasonably comfortable, even in the coldest weather. Oh, and insulated knee-high snow boots as well. I pushed wearing my regular boots as long as possible, then realized that I felt colder because wind could blow up the cuffs of my pants. Pulled on the insulated snow boots and did that ever make a difference.

The kind of riding I'm doing now has changed. Summer was a focus on schooling and refining skills--for both of us. I hadn't realized until this summer just how much I had modified my riding to accommodate Mocha's needs, then switched to young horse schooling with Marker. I've been doing light weight work to deal with tight back muscles, and that's also helped as well. But I needed to work on bringing my legs back, which...those tight muscles had been impairing. One of those sneaky impairments that creeps up on you with age, I suppose.

In any case, Marker and I spent a lot of time in the arena this summer, and it's pretty much paid off. He hits his canter leads darn near 99% of the time. He's much more confident and strong when cantering, too. I've found that some horses really do just need to have a lot of time cantering to be strong enough under saddle to be relaxed about it, and given that a lot of gaited horses often struggle with cantering under saddle, we had to spend a bit of time conditioning. And it's not consistent yet. I have a lovely, balanced, rocking-horse left lead canter. Right lead? He still wants to rush and speed up. More work and conditioning is required--rushing is a sign of tension. It'll happen over time.

But winter work is different from the intensive schooling of summer. Oh, I do a little bit of schooling. Right now I'm working on developing seat cues, and he's picking it up pretty well. We're doing small serpentines and circles where the primary cues are the weight of the outside seatbone and the turn of my head, and that's coming along nicely. It's interesting, because Mocha was the cattier of the two horses--up to her last days, she was capable of executing a sharp 180 turn off of her haunches tighter and faster than most of the other horses (it was one of her evasions when another horse started making her move her feet). However, she wasn't as responsive to seat cueing as Marker is so far--most of the time, I had to tune her up before I could casually weight a seatbone, turn my head, and have her respond. Marker? He has picked it up quickly. No tuning required. I can get that response with a weighted seatbone and head turn early in the ride, rather than later, after starting with heavier cues, then softening to the lighter cues. It'll be interesting to see how light I can make him on a consistent basis.

I also invested in some inexpensive oversized stirrups because I didn't like the way my regular stirrups work with the winter boots. Additionally, I had this sneaking hunch that they were making it more difficult to get on from the ground. Well, I've only had two sessions with the new stirrups but my theory has been confirmed--it is easier to get on with the oversized stirrups (just a hair wider), and they work a lot better with the winter boots. Plus they hang better from the fenders than the originals did. The biggest drawback is that they are plastic, so I wonder how durable they'll be in cold weather. Oh well. Gives me data should I want to upgrade to more expensive ones.

But, mainly, winter riding is just more about keeping up the fitness for both of us. A slower, more relaxed riding time. A chance to work on our connection with each other. A throwback to the days of my youth, only with a better quality of horse.

If I had been told years ago that I would still be throwing a leg over the back of a good horse at age sixty-eight, I might not have believed it then. Now? Well, I'm gonna do my darnedest to keep going as long as I can.

And in order to do that, I need to drag my rear out to the field and ride the horse in all seasons.

S	M	T	W	T	F	S
						1
2	3	4	5	6	7	8
9	10	11	12	13	14	15
16	17	18	19	20	21	22
23	24	25	26	27	28	29
30

Diary of a Mad Webfarmer

Reading

SARAH STEPHENS IS NOT YOUR AI GRANDMA

Winter Ramblings with Horse

Profile

April 2017

Most Popular Tags

Page Summary

Style Credit

Expand Cut Tags