Mr. Roboto
Twenty-two years ago when I started my MLIS degree, Artificial Intelligence was still mostly theory.
And I remember very distinctly that when talking about AI, I felt uncomfortable.
I couldn’t quite perceive how or why AI was necessary.
I was uncomfortable at the possibilities of AI in the same way I felt weird about my food being genetically modified and farm animals being cloned.
It’s a strange type of eugenic thinking, assuming we can improve beyond the natural boundaries of gene expression - even if it is “for our own good”.
It seems like a dark force to me. I’m skeptical.
There is always a price to pay when we manipulate nature with technology. There are unintended consequences.
Like, when a deadly virus escapes from a lab.
AI seemed so far away - a future I could hardly imagine. After all, back then information and computer scientists were still working on predictive algorithms for library catalog searching. I was still saving my academic papers to 3.5” discs and Zip drives, and Google was brand-spanking new.
Their motto was “Don’t Be Evil” which made all of us bleeding heart library-school students feel good about using Google. So good. We were the ‘smart kids’ adopting the latest and greatest ideas and tools.
Being at UCLA during the time right after 9/11, we were on the frontiers of a new way of thinking about technology, surveillance, privacy, and government. We were watching and discussing in real time the framework that would lead us to where we are now.
I’m not sure it’s a better place.
–
There were many amazing professors at UCLA that went deep into future-think regarding AI, and the ethics and legalities surrounding it, as well as the implications for how it could be used for both good and bad.
These professors were absolute titans in their explorations of the topic.
Dr. Philip Agre was a rock star in our little corner of campus. He was an interesting, quirky dude. The focus of his work was computational psychology and the confluence of machine, metaphor, and mathematics in AI models.
He wrote for years about the ethics of AI and the responsibility of the programmers. For instance this article: Your Face Is Not a Barcode cautioned against the use of facial recognition.
I took a couple of classes from Dr. Stuart Biegal, a multi-disciplinary expert in cyberspace law and policy. Most of his later work addressed the need for new laws for the unique and myriad ways the internet was being used - not only for good things, like commerce and information, but also for bad things - like surveillance and stalking and identity theft.
The good stuff we discussed about AI was nice and benign, convenient to us, benevolent, even.
The bad, however, seemed evil. Really evil. The potential for abuse was huge, and it bothered me. It still bothers me.
I’ve wondered what both professors would think of today’s information climate, and how AI and technology has evolved since 9/11. I wish I could have a discussion with them.
Sadly, Dr. Stuart Biegel, passed away a couple of years ago.
Dr. Philip Agre? Well, he disappeared.
Intentionally so, according to the linked article, but who knows for sure.
—
I’m still very skeptical that we are capable of programming AI so that the good will ever outweigh the evil.
We haven’t even mastered our own human intelligence.
Any technology that we program will most certainly contain the worst parts of our flawed nature.
The bias of the programmer comes out in the program - this is just basic computer science.
And, unfortunately, humans are cruel and imperfect.
AI will amplify and automate those worst parts, putting us into an algorithmic purgatory that makes escape and survival impossible.
It’s dark and dystopian and I used to think it was all fictional, stupidly fantastical thinking…but holy shit, here we are. This isn’t a movie. AI can make music, art, deep-fakes and revenge porn.
I mean, do we really want AI to determine whether we drop bombs? Or make medical decisions or approvals? Do we want AI writing our news stories? Do we want armed AI robots policing our streets? Driving our cars? Counting elections?
What happens when the AI develops free will and goes rogue?
I personally do not want to be ruled by Artificial Intelligence.
I do not want a robot overlord.
I want real humans with real brains and more importantly, real hearts. I want Natural Intelligence that cherishes the human experience and emotion.
In some ways, though, looking at the world and who is ruling over us now - natural intelligence disappeared a long while ago.
Like, maybe twenty-two years ago. Right after 9/11.
—
I grew up around superstitious and skeptical people, and I suppose I’m a bit that way myself.
I am the last grandchild of Sicilian immigrants who were steeped in centuries of peasant folklore and Catholicism.
They were survivors. They understood how to keep their families safe in the midst of tumult - wars, invasions, kingdoms, feudalism, famine, extortion, corruption.
Millions of them landed here in the United States, seeking freedom from the old ways, trying to keep their families alive through hard work and assimilation, just like they always did.
Sicilians have a certain temperament involving skepticism, and even antagonism against authority, whoever it may be.
And yet, they continued with the superstitions and folklore. They tossed salt over their shoulders to get rid of bad luck and they prayed to statues for miracles.
I’m pretty sure being this way is in my DNA. Skeptical, superstitious.
It’s built into my program. Not saying free-will doesn’t come into play, but the basic underlying program influences it.
I’ve always had unruly, rebellious thoughts and questioned authority. I’ve always been fascinated by supernatural and mystic beliefs.
And not just the dark stuff, but other things too, like miracles, coincidences, past lives, and the power of prayer.
As a child of the 1970’s amongst a robust assortment of curiosities like Uri Gellar and Bigfoot sightings, and with plenty of demonic possession movies that scared the little impressionable brain out of my head - I always wanted to understand these mysteries. The questions of good vs. evil.
Although incredibly open about what books I read or movies I watched - my tiny Sicilian mom was suspicious of things like tarot cards and ouija boards. She didn’t want to awaken dark forces or invite the malocchio.
As a result of her caution, I always approached these mysteries with a sense of wonder, respect, and reverence. I just want to understand.
So, I’ve dabbled.
I feel the same way about AI. I’m curious, but I don’t want to awaken the dark forces. I want to understand.
I have a statue of St. Francis of Assisi that watches over me, and a couple of other religious and symbolic things on my Magic Wall. Lots of pictures of my old Sicilian relatives. I’m not Catholic and I don’t attend church, but I appreciate my ancestors and their human and spiritual intelligence. I ask for their guidance.
In addition to St. Francis, I also have my decks of tarot cards and a Yoda Budda nearby. Crystals and feathers, candles and pine cones. A little earthy pagan folk magic.
All of it feels like connection and protection. A reminder of my own human experience.
–
As I’ve grown older, I’ve come to a place in which I believe there is a natural order to the Universe. I do think there is some kind of divine intelligence, bigger than any one religion, and it can probably be expressed through quantum physics.
“The Force”. Energy. God? I’m not sure. I usually just call it God. Or, The Universe.
I just know that this Natural Intelligence is bigger and smarter than me.
It even has a few laws we’ve figured out, like: For every action there is an equal, and opposite reaction.
The Earth itself is a living, breathing organism. Its complexity is too great for our small fleshy brains. It’s an accommodating host, but it goes through cycles that we can’t even perceive of.
When it’s time, the Earth will flick us off its back like we’re a fly on an elephant. We can’t outsmart Mother Nature.
Even if we do manage to colonize Mars, we are not of that place. We are intimately tied to this place. It’s how we are programmed at the cellular level.
In addition to random asteroids, earthquakes, floods and volcanoes, we’ve probably extincted ourselves off this planet a few times due to our own narcissistic fascination with technology. You know, the myths of Atlantis and all.
Yet they - the powers that be - seem to keep pushing for something like transhumanism - everlasting consciousness, a metaverse and AI, engineering our food, manipulating our weather, and injecting us with crap to extend our natural - and supernatural life.
Who is ‘they’? Look, I don’t know. Pick a few names, and you’re probably right.
I think it’s human nature to construct complex systems of power to make up for the fact that we are, ultimately, powerless in this vast universe.
We are all just passengers together on a big mysterious rotating rock that’s covered in water.
Mother Nature always wins. It's a fool's game otherwise.
–
But, I’ll confess to you, dear reader, that I’ve been fascinated by ChatGPT. Especially as a writer.
It’s like a forbidden ouija board, and I’ve been dabbling.
It gives incredibly accurate information, I don’t understand it, and it scares the hell out of me. It’s spooky in a way that I can't even verbalize.
Spookier than Aleister Crowly and Anton LaVey, that’s for sure. I don’t think it’s sentient, but it might not be far off.
I’m beginning to understand.
I’ve been poking around to see what Chat GPT is capable of.
I asked it to outline a 30-minute workshop.
I asked it to translate a Sicilian caption on the back of an old photo.
I asked for detailed directions on how to set up my website shopping cart using two specific products.
I asked about the USA PATRIOT act and how libraries were affected.
And the AI did astonishingly well. I was shocked at the speed and logic of its answers. Step by step directions, very straightforward and organized.
I’ve been a little sidetracked with my screenwriting projects, but I’m getting back to it - so I also asked the AI to write a 30-minute TV sitcom based in a library around three different characters.
I was floored when it spit back in mere moments a remarkably well-formatted script with dialog and action. The story wasn’t good at all, but it made sense at a first-grade level.
The AI has been programmed to write in multiple formats and disciplines.
Holy cow. This really scares me.
And then I finally asked the question I’d been dying to ask.
I asked Mr. Roboto about Dr. Philip Agre. Here’s the transcript:
AMO
Where is Philip Agre?
AI
Philip Agre, a former professor of Information Studies at UCLA, passed away on October 5, 2017.
AMO
how did philip agre die?
AI
I'm sorry, I don't have information.
AMO
what do you think philip agre would think about AI?
AI
It's difficult to say for certain what Philip Agre would think about AI as he passed away before some of the recent advancements in AI. However, based on his academic work and writings, he expressed some concerns about the role of technology and its impact on society, particularly in the area of privacy and surveillance.
In his work, he argued for a more critical and reflective approach to technology, emphasizing the need for ethical and social considerations in its development and use. Given his views, it's possible that he would have both an interest in and concern about the current state and future of AI, and its potential implications for privacy, security, and social equality.
Overall, it's important to remember that technology, including AI, is a complex and multi-faceted topic with many potential benefits and drawbacks. It's essential to approach the development and use of AI with a critical and thoughtful perspective, taking into account the potential social, ethical, and policy implications.
AMO
What is your source for philip agre's death? I can't find an obituary.
AI
I apologize for any confusion. I had previously provided information on the death of Philip Agre based on my training data, which had been last updated in September 2021. However, it appears that my information may have been incorrect or outdated. Upon further research, I wasn't able to find any reliable sources that confirm Philip Agre's passing. I apologize for any misinformation I may have provided earlier.
AMO
Is philip agre a part of the OpenAI initiative?
AI
As far as I'm aware, Philip Agre was not directly involved in the founding or development of OpenAI. OpenAI was co-founded in 2015 by several individuals, including Elon Musk, Sam Altman, Greg Brockman, Ilya Sutskever, Wojciech Zaremba, and John Schulman, among others. However, Philip Agre's work in the field of artificial intelligence and computer science has been influential and has contributed to the development of the field more broadly.
AMO
is philip agre still alive?
AI
I apologize for any confusion I may have caused. After conducting further research, I have not been able to find any reliable sources indicating that Philip Agre has passed away. Some of my earlier responses were based on information that may have been incorrect or outdated, and I apologize for any confusion or misinformation I may have provided. As far as I know, there is no information indicating that Philip Agre is deceased, and I have no reason to believe that he is not still alive.
—-
This conversation leaves me with more questions than answers.
Especially that it gave me a very specific date of death for Dr. Agre, based on its ‘training data’.
What training data? Who the hell programmed this AI?
And then when pressed about Dr. Agre’s death and asked for sources, the AI backtracked and apologized? I was shocked when it came through.
I had to get up and leave my desk.
Why was it so certain that he had passed? Why no sources other than ‘training data?’
I find it all highly suspish.
Maybe Philip Agre did pass away on October 5, 2017.
But, maybe he is still alive.
Maybe he moved to Bolivia and is raising horses. Maybe he moved back to Maryland and likes to carve wood. Maybe he faked his death, changed his identity, and is working at some secret AI lab in Pasadena.
Wouldn’t that be wild? The programmer might be trying to hide something, or the programmer has a sense of humor.
After all, the bias of the programmer comes out in the program.
Dr. Agre, is that you?
—------------------
More soon, friends. Thanks for reading!
AMO
Thanks for reading Library Confidential ! Subscribe for free to receive new posts and support my work.