Published using Google Docs
AI in Science Fiction
Updated automatically every 5 minutes

AI in Science Fiction- a history of humans imagining smart machines

The small talk routine -- > https://www.youtube.com/watch?v=9FqFm_vmVnE 

Okay we're kind of caught up to now on the internet! We skipped the last 15 years or so of internet developments, but that is A) the era that most of us in this room directly lived through online and B) is kind of the general context for the materials of this class. So we’ll be approaching 2005-2019 in computers in general from here on out.

But before we switch tracks fully into the contemporary computing world, I want to jump sideways a little and talk about science fiction, which has profoundly influenced many of the designs and ideas of our technological landscape.

The other thing we’re picking up this week is the idea of artificial intelligence, which hasn’t been a big part of our conversations about desktops, file systems, the browser, early internet structures, or the legal clauses and legislation that keeps all that stuff running.

This class is really about data, how data is constructed, gathered, and used. And the number 1 hot data usage in 2019 (maybe outside of advertisement) is artificial intelligence and machine learning applications. We aren’t going to be directly covering any machine learning tools in this course (although I can point you at further resources if you are interested) but a lot of the datasets we will use and interact with will have been made directly for the training of machine learning, so it’s important that we have a very high-level understanding of how they work, and why data for them gets formatted in the ways it does.

(Gonna switch over to a few “here’s the intro to the idea” Youtube vids real quick)

2 minute explanation of machine learning → https://youtu.be/QghjaS0WQQU 

Basically, machine learning is the general term for when computers learn from data. There are many different ways that machines can learn (these are what folks in this industry mean when they say "algorithms"). These algorithms are generally either supervised or unsupervised.

When you make data for a machine learning algorithm, it can be in a input-output pair (for example, a picture of a banana and the label “banana”, as well as a picture of an apple and the label “apple”), or just have the inputs (pictures of bananas and apples). Supervised learning algorithms require input-output pairs, while unsupervised learning requires only the input data.

Supervised algorithms:

You feed it an example input (picture of banana), then the associated output (“banana”). Also, picture of apple, “apple”. You repeat this many many times. Eventually, the algorithm picks up a pattern between the inputs and outputs. Now, you can feed it a brand new input, and it will predict the output for you. Is a mango more banana or more apple? We can now find out.

Unsupervised algorithms:

You feed it an example input, without the associated output (just photos of apples and bananas). You repeat this many many times. Eventually, the algorithm clusters your inputs into groups. Now, you can feed it a brand new input, and the algorithm will predict which cluster it belongs with. Again, it will send our mango to one of the two groups.

(This supervised/unsupervised example was edited down from the world’s only helpful Youtube comment, by Abdullah Aghazadah.)

Brief history of AI video --> https://youtu.be/ikzJjPYGc8U 

(Sorry about the kind of silly “oo the singularity” overview but it does a pretty good job covering the baseline of stuff we wanna know going in~)

Machine learning is not exactly AI (and vice versa) but they are related concepts that get used interchangeably in the news in particular. The biggest difference between them is that AI is the practice of making machines smart in ways that look like human cognition (even if the goal isn’t a sense of self or consciousness). Machine learning is a current implementation towards the goal of AI.

Some version of the idea of artificial intelligence has been around basically forever. It’s a really old story with humans, although past visions look a little different than what we think of as AI now.

from https://emilyflorenceblog.wordpress.com/2017/10/07/riveting-ramblings-myth-market-golem/

On Golems:

Golems are an old idea, with has been a part of Jewish folk belief and practice since early Judaism. They basically stem from the idea that if thinking humans came from mud via the will and word of divinity, then it may be possible to animate mud in the same way. Golems are animated sculptures of clay, baked in ovens and given will. However, no golems are fully human - they are often personified as being extremely strong but cognitively or emotionally limited, or lacking the ability to speak.

It was believed that golems could be activated by the ritualistic use of various letters of the Hebrew Alphabet, which formed a "shem". The shem was written on a piece of paper and inserted in the mouth or in the forehead of the golem, animating it with divine instructions.

There are stories about these words being ‘edited’ to change the behavior of the golem - one such story describes a golem with emet (אמת, "truth" in Hebrew) written on its forehead. The golem could then be deactivated by removing the aleph (א) in emet, changing the inscription from "truth" to "death" (met מת, meaning "dead").

Golems are old stories, but weren’t a part of popular Jewish culture until the 17th century, in Prague. The Golem of Prague was animated as a defense force- an unkillable creature of clay to defend the Jewish community against anti-semetic attacks. The Golem was called Josef, and it was claimed he could make himself invisible and summon dead spirits. The story goes that eventually the Golem grows wild (for various reasons), and is turned into dust on the steps of the Old New Synagogue, where it is stored in pieces for when it is needed again.

This story was popularized by German Jewish writers and poets in the 1800s, and it entered popular culture as a romantic-era conceit. The world ‘golem’ appears once in the bible, where it denotes an unformed, embryonic Adam- the form of a human before being given a soul. 19th century writers and poets developed the idea of the Golem as doppelgänger.

The curator Emily D. Bilski has this to say about the relationship between Golems and AI;

“The motivations and methods for creating a golem, as well as the golem’s actions once it has been brought to life, have provided powerful analogies for the potential risks and benefits of creating artificial intelligence.

"Each of the many manifestations of the golem – in literature, theatre, film, visual art, or popular culture – has reflected the concerns of their creators and the anxieties of the times in which they lived, whether the increased brutality of technological warfare, political movements run amok, the threat of nuclear annihilation, or scientific advances in biotechnology, AI, and big data. Both the golem story and Frankenstein have been understood as expressing the ‘horrors’ that science and technology can wreak.

Yet it is the golem that underscores the ambivalent relationship we have with scientific and technological progress.

In this regard, the scientist seeks to benefit humanity, just as the goal of the rabbis was to rescue Jews threatened by violence; therefore, any resulting danger is an unintended consequence. It is this dialectical nature of the golem – at once saviour and destroyer – that offers fertile comparisons with AI, and has inspired writers, filmmakers, and artists to interpret this archetypical."

In 1965, the Weizmann Institute in Israel built a new computer and at the dedication ceremony Gershom Scholem (a Jewish scholar) said that it should be named ‘Golem Aleph’ [Golem Number One], concluding:

"So I resign myself and say to the Golem and its creator: develop peacefully and don’t destroy the world. Shalom."

Others have made parallels with other historical belief structures, like the golden robots of the Greek god Hephaestus, who created intelligent robot servants for his home, as well as two mechanical, immortal dogs out of gold and silver in order to guard the palace of Alcinous, automatic tables that obeyed the voices of the gods, who called the tables to provide them with food and drink, two wild bulls with bronze feet, with voices coming out of their nostrils and fire coming out of their mouths, and the giant sentry of Talos, whom he gave to King Minos in order to guard the island of Crete and oversee the implementation of laws.

Jason and the Argonauts, 1963

Talos had some pretty brutal abilities, like the capacity to heat its body to red-hot, then hug dissidents to death. It is described has having “liquid metal” lifeblood, than run inside itself and was plugged at the foot. The sorcerer Medea convinced Talos that it could become mortal if it unstopped the plug, at which point it died.

A dramatization → https://youtu.be/fdgY9vgRcRw 

The Greek myth of Pygmalion and Galatea also pokes at this idea, of a sculpture so lifelike and detailed in its construction that it becomes alive.

Others have pointed to Shinto belief systems as a useful touchstone for robotics and AI, as intelligence and presence isn’t strictly limited to living creatures, but is rather distributed through a larger system of things that have their own logical selfhood.

Formal logic and philosophy:

The development of ideas of formal logic in general had a big impact on ideas of intelligence and what it meant to reason. Formal logic has been around for a few thousand years, but the touchstone I want to look at briefly is the Catalan / Spanish philosopher Raymond Llull, who lived from 1232–1315.

Llull developed several ‘logical machines’ which could combine basic and undeniable truths by simple logical operations, producing new mechanical meanings- in such ways “as to produce all the possible knowledge”.

 

The writer Jorge Luis Borges describes one in 1937;

It is a schema or diagram of the attributes of God. The letter A, at the center, signifies the Lord. Along the circumference, the letter B stands for goodness, C for greatness, D for eternity, E for power, F for wisdom, G for volition, H for virtue, I for truth, and K for glory. The nine letters are equidistant from the center, and each is joined to all the others by chords or diagonal lines. The first of these features means that all of these attributes are inherent; the second, that they are systematically interrelated in such a way as to affirm, with impeccable orthodoxy, that glory is eternal or that eternity is glorious; that power is true, glorious, good, great, eternal, powerful, wise, free and virtuous, or benevolently great, greatly eternal, eternally powerful, powerfully wise, wisely free, freely virtuous, virtuously truthful, etc., etc.

I want my readers to grasp the full magnitude of this etcetera. Suffice it to say that it embraces a number of combinations far greater than this page can record. The fact that they are all entirely futile—the fact that, for us, to say that glory is eternal is as rigorously null and void as to say that eternity is glorious—is of only secondary interest. This motionless diagram, with its nine capital letters distributed among nine compartments and linked by a star and some polygons, is already a thinking machine. It was natural for its inventor—a man, we must not forget, of the thirteenth century—to feed it with a subject matter that now strikes us as unrewarding. We now know that the concepts of goodness, greatness, wisdom, power, and glory are incapable of engendering an appreciable revelation. We (who are basically no less naive than Llull) would load the machine differently, no doubt with the words Entropy, Time, Electrons, Potential Energy, Fourth Dimension, Relativity, Protons, Einstein. Or with Surplus Value, Proletariat, Capitalism, Class Struggle, Dialectical Materialism, Engels.

The idea that the mind and body could be separate entities gained a lot of traction in the 1700s-1900s, and basically all popular Western philosophy from that era stems from that idea (which was pretty radical at the time! Also wrong! Kind of? But yeah, complicated!)

You can imagine that being able to pull apart concepts of a mental self from an embodied self meant new avenues of thought, in which calculation and thought could happen without a body. These ideas along with new developments in medical science are what form the backdrop to Mary Shelley’s Frankenstein. Obviously, Frankenstein’s Creature isn’t a synthetic creature, but it is a new intelligence in a reanimated body, who develops into a person of revenge and violence out of neglect and abuse from society at large. Worthy concerns to carry forward for sure.

From here on out science fiction in a genre is in full swing and we get an android or clockwork being or something every couple of years! The 1921, the term ‘robot’ is introduced by Czech playwright Karel Capek in his play “R.U.R.” (Rossumovi Univerzální Roboti (Rossum's Universal Robots)).

Three robots in R.U.R., wikimedia commons

R.U.R. is set in a factory that makes artificial helper people. They are actually made of flesh, and are just synthetic humans- but are trained to be a servant class. They eventually rebel, leading to the extinction of the human race. It was a huge hit, wildly.

In 1927, Metropolis was released which both depicted the robotic pace of human life in the future city, as well as one of the first on-screen robots, the Maschinenmensch. This robot houses the intelligence of the human Maria, who is transferred to an artificial body and brain.

-- > https://youtu.be/IcReykfvqi4 

Here is another clip of Data that I couldn’t find another spot for. Idk it seemed relevant to our goals so here it is- > https://www.youtube.com/watch?v=gCOFtE4j41A 

A copper engraving of the Turk, Wikimedia Commons

Okay, so the story that I want to end on is about “The Mechanical Turk”, a chess-playing automation that shares a name with the Amazon micro-tasking service (which honestly is a bit dire?).

https://www.mturk.com/

https://www.reddit.com/r/mturk/ 

Automatons were fashionable all through this era, and the best of them were beautifully created inventions. They were kind of a way for clockmakers and machinists to show off.

From the 1770s, this automaton called The Writer, was designed and built by Swiss watchmaker Pierre Jaquet-Droz. Inside the boy is 6,000 custom made components that create a fully self-contained re programmable writing machine.

Or this remarkable silver swan -- > https://youtu.be/whzoMIL-y3k 

The Mechanical Turk though was an automaton constructed in the late 18th century. The turk was carved of wood, dressed in mystic robes and seated at a low table. He was capable of playing (and winning) at chess, as well as completing the puzzle game ‘the knight’s tour’, in which a knight occupies every square of a chess board exactly once, visiting all 64 squares. The table and Turk contained many doors, in which an audience may see the complicated clockwork that drove the automaton. It was nothing short of a miracle.

Video on the chess turk.→  https://www.youtube.com/watch?v=0DbJUTsUwZE 

The Turk was fascinating because it claimed a middle-space between ‘person’ and ‘object’; a theoretical bot so good that it transcended mere routine and gained a shine of real intelligence. And automata of the era was really great! Miraculous even! But not quite, this miraculous. Its very human that we find a machine performing the actions of a human chess master more compelling than the human chess master- but there we are.

This is in a lot of ways a perfect metaphor for where we are at in general with AI, because although our systems do remarkable things for us all the time, it is just as often a layer or trick on top of some person, somewhere.

Next time we are going to talk about how the desire for human-like AI has resulted in some really weird situations of human people performing as virtual assistants / robots / chatbots. It’s still cheaper to hire outsourced labor, but it’s good optics to have smart machines. The reading is related to this topic!

I wanted to cover this history of science fiction and AI because we’ve been performing as machine intelligences for a long time, and its influenced how we think they should behave. They’ve bled into each other.

-- > replicant test scene, bladerunner. https://youtu.be/yWPyRSURYFQ 

Reading for Monday, Sept 16 -

Assignment for Wednesday, Sept 18 -

Imaginary machine

Construct a human-made dataset, artwork, poem, story, video, or other creative project that is posing as a thing made by machine. Over the weekend, begin to think about the type of language and selfhood that is manifested online that marks the difference between humanity and bots. Come with an idea of the type of thing you want to make! (This is a small project - no need to go overboard. )