Learning Curve: Living (or not) with Aliens

Trade out the word “Artificial” for “Alien” and AI starts to make more sense. There is nothing fake about the intelligence, although there is a kind of alchemy when a Large Language Model (LLM)—the “machine learning” on which an AI is based—coalesces into an algorithm that can write press releases, design proteins, generate art and pass the bar exam without skipping a beat. AI can think so much faster than we can. It can also “think different,” processing information in ways we simply aren’t built for.

“Unsupervised” is a stunning exhibit of work by Refik Anadol at the Museum of Modern Art (MoMA) that embraces and explores both AI’s speed and its alien ways.

The title is in counterpoint to “supervised learning,” which is how AIs are typically trained, using datasets that humans have tagged. The floor-to-ceiling videos in the exhibit are generated by an algorithm that has assimilated the museum’s entire collection: 138,000 images. Anadol instructed the AI to tag and organize the massive dataset using whatever criteria made the most sense to it.

A 3D computer map revealed large blank areas between clusters of artwork, reminiscent of galactic dark matter. Anadol asked the program to fill in the blanks. The result is imagery that is dream-like yet specific, hallucinatory yet intentional. There is undeniable intelligence at work, a logic behind every move.

“The AI is not using our value system. It’s not using our points of reference. It’s…peering into another type of mind,” observes MoMA curator, Paola Antonelli.

•••••••••••••••••••••••••••••

Of course, that’s not the ChatGPT version of AI, which is deeply rooted in the fraught terrain of the “supervised.”

It still strikes me as preposterous that we’re assuming that a single image can be given a label, a single word, when we know about the multiplicity and complexity of a single image. The idea that we can so benignly label something as a chair, and a person as a debtor or a kleptomaniac—these are things that are literally happening today in datasets. And the risk here is that we’re starting to see a very simplified and, in some ways, really bleached version of the world.” Kate Crawford, author, The Atlas of AI

ChatGPT “writes” by predicting the next word in a sentence. The algorithm literally has no skin in the game. There are no consequences to a computer program that spits out a lie. It has no need to understand what it’s writing.

But we do. We need to know what true and what’s not.

Writing, which so many are eager to outsource, is a process that helps clarify thought. To delegate that is give the away the keys to the kingdom.

In the beginning was the word. Language is the operating system of human culture. From language emerges myth and law, gods and money, art and science, friendships and nations and computer code. A.I.’s new mastery of language means it can now hack and manipulate the operating system of civilization. By gaining mastery of language, A.I. is seizing the master key to civilization, from bank vaults to holy sepulchers….

…A.I. could rapidly eat the whole of human culture — everything we have produced over thousands of years — digest it and begin to gush out a flood of new cultural artifacts. Not just school essays but also political speeches, ideological manifestos, holy books for new cults. By 2028, the U.S. presidential race might no longer be run by humans. — Yuval Harari, Tristan Harris and Aza Raskin, NYT

It is still early days in our real-time experiment of living with an Alien Intelligence.

To harness it to write copy that no one should be writing and no one should ever have to read seems like a low bar.

But an Alien Intelligence that could help navigate through the many knotty problems humans have created could be welcome. Tipping points are colliding. Time is running out. A new way of thinking, preferably one where hallucinations have been replaced by ethics, could make a difference.

Or it could be the kind difference that puts all life in the cross-hairs.

…To visualize a hostile superhuman AI, don’t imagine a lifeless book-smart thinker dwelling inside the internet and sending ill-intentioned emails. Visualize an entire alien civilization, thinking at millions of times human speeds, initially confined to computers — in a world of creatures that are, from its perspective, very stupid and very slow. A sufficiently intelligent AI won’t stay confined to computers for long. In today’s world you can email DNA strings to laboratories that will produce proteins on demand, allowing an AI initially confined to the internet to build artificial life forms or bootstrap straight to postbiological molecular manufacturing.

If somebody builds a too-powerful AI, under present conditions, I expect that every single member of the human species and all biological life on Earth dies shortly thereafter… — Eliezer Yudkowsky, Machine Learning Research Institute, Time magazine

So Alien Intelligence as the latest iteration of Frankenstein’s Monster, with a Changeling’s capacity for mimicry, the imagination of an artist and sociopath’s complete lack of remorse.

Great.

Of course, all this speedy thinking requires vast amounts of energy. By 2030, which is only seven years off (so the time it takes to grow a second-grader), computation will account for nearly a third of all electricity consumption worldwide. Ten years after that, we won’t be able to generate enough power to satisfy AI’s hunger. So now it’s the Borg Queen looking to plug into whatever she can.

Researchers are working greener, energy miser chips, but like climate change, it’s a race where we seem to be falling further behind each day.

That leaves The Matrix.

The Matrix?

Who would have thought that using AI for writing Instagram captions could be a vision of the good old days?