• Rox's Picks
  • Posts
  • AI in 2023: What the hell can we do? (Part 2) | RP 85

AI in 2023: What the hell can we do? (Part 2) | RP 85

Generative AI in creative work & what teachers can do about ChatGPT

Hey friends!

Last week’s newsletter had a 38% open rate. The top link you clicked on was Wait But Why’s article on “The AI Revolution: The Road to Superintelligence.”

Last week, I shared a couple of articles that give a broad sweep of the AI landscape thus far. Today, I’m sharing a few ideas on how AI affects how we work and learn to wrap up this two-part series on AI.

This week’s articles shed a lot of light on how AI affects our daily lives in the near term, particularly in 2023 and in the next 2 years:

  1. "Generative AI: Autocomplete for Everything” by Noah Smith and roon

  2. “AI Homework” by Ben Thompson

With that, here’s what I learned, shared, and paid attention to this week about AI:

1. “Will AI take my job?”

Recently I've been seeing more artists publish posts that disown AI:

The captions say something along the lines of, "No AI was used in the creation of this work. I do not allow the use of my work in AI generators."

I’m not invalidating how an artist decides to share her work, but this week’s article from Noah Smith and roon offers an alternative perspective on the use of AI in creative work. From the article (emphasis mine):

"We think that the work that generative AI does will basically be 'autocomplete for everything'… Midjourney isn’t dreaming up 'Guernica' – there’s a limit of narrative abstraction to the text prompt after which it loses its grip and starts spitting out nonsense. It’s up to the artist to distill a creative motive, an inspiration, into simpler ideas a machine can understand, and iterate on this. At the moment, prompting is an inexact science where a human must find the right way to communicate intent to an AI – a problem not dissimilar to difficulties of delegating to other humans! Then finally, the artist must fix up any details the AI may have screwed up."

In other words, what if we embraced our new AI underlings as creative collaborators instead of as artistic rivals who are out to take our jobs?

This isn't a new suggestion. Creative collaboration has been a celebrated and crucial part of the creative process for hundreds of years.

Take Andrea del Verrocchio, a sculptor and painter who ran an important workshop in 15th century Florence. Even though his apprenticeship produced a handful of important Renaissance artists, Verrocchio’s workshop was less like an art school or studio, and more like a commercial design shop that produced paintings and sculptures on demand.

In workshops like Verrocchio’s, artworks were products, not expressions of individual creativity. Like many master artists of his day, Verrochhio employed studio assistants called garzones (a.k.a. live-in interns) with whom he collaborated to create the commissioned works. One author writes, "The goal was to produce a constant flow of marketable art and artifacts rather than nurture creative geniuses yearning to find outlets for their originality."

Verrocchio was best known for his sculpting, not his painting skills. Collaborating with talented young artists was crucial for business. As a result, even an untrained eye can see the marks of this collaborative process in pieces like Verrocchio’s The Baptism of Christ.

If you zoom in on the two angels, the left one in particular possesses what experts call "a melting softness." This is in contrast to the harsh sculpture-like angularity that pervades most of the work — a calling card to Verrocchio, the sculptor.

Even though this painting is attributed to Andrea del Verrocchio, it features contributions from apprentices in his studio. One of these apprentices, a then-unknown 17-year-old painted the angel on the left. This student would soon eclipse his master in skill, fame, and historical notoriety: Leonardo da Vinci.

Walter Isaacson writes about this master-apprentice collaboration in his book on the artist:

"X-ray analysis shows that Verrocchio, with his lesser feel for nature, had originally begun the background by drawing a few rounded clumps of trees and bushes, more wooden than sylvan. When Leonardo took over, he used oils to paint a richly natural view of a languid but sparkling river flowing through rocky cliffs, echoing his Arno River drawing and foreshadowing the *Mona Lisa*. Other than Verrocchio’s pedestrian palm tree, the backdrop displays a magical mix of natural realism and creative fantasia."

Of course, working with AI is different from working with other human artists.

I'm not here to debate the ethical implications of attribution and of using generative AI to produce creative work — personally, I’ll never publish anything that I copy-pasted from a ChatGPT response — but I am curious on how creatives can use AI as a way to jumpstart our creative processes and help us produce better work sustainably.

A final word on this from Smith and roon:

“Just as some modern sculptors use machine tools, and some modern artists use 3D rendering software, we think that some of the creators of the future will learn to see generative AI as just another tool – something that enhances creativity by freeing up human beings to think about different aspects of the creation.”

2. “What should schools do about AI?”

“AI Homework” by Ben Thompson

In this article, Thompson likens the use of ChatGPT in schools to the use of the calculator. He says:

"The obvious analogy to what ChatGPT means for homework is the calculator: instead of doing tedious math calculations students could simply punch in the relevant numbers and get the right answer, every time."

How teachers adapted to the calculator is a good starting point for dealing with ChatGPT in classrooms. Instead of banning calculators from classrooms, teachers asked students to show their work.

Instead of banning ChatGPT, educators can teach students to think critically. AI has devalued the skill of regurgitating facts on demand. Educators will now be forced to adapt teaching methods to emphasize asking questions thoughtfully, thinking critically, and editing ruthlessly.

With the advent of generative AI, educators and institutions must accept:

  1. There will be a lot more authoritative misinformation coming at internet users in the next few years. This means that…

  2. Editing and critical thinking skills will become even more valuable.

Thompson concludes:

"There is a lot more bad information for the simple reason that it is cheaper to generate. Now the deluge of information is going to become even greater thanks to AI, and while it will often be true, it will sometimes be wrong, and it will be important for individuals to figure out which is which."

AI’s strong suit is probabilistically remixing available content on the internet. Humans’ strong suit is giving our AI underlings the right prompt, then verifying and editing the output. To paraphrase Thompson, while calculators offer an answer we can trust, generative AI offers a starting point we can correct.

What’s next?

This newsletters wraps up my survey of the AI landscape. In the last 2 weeks, I realize I’m less interested in Ray Kurzweil’s singularity — the point of no return when AI surpasses human intelligence — and on whether superintelligence is going to take over the world, than I am in what we can do about it in the near term.

I’m psyched to dive into a different part of the tech and business industry next week.

I’m finishing up my read of Sapiens and will also be publishing an essay on that this month. Stay tuned. 🫡

Stay strong, stay kind, stay human.

Till next week,

roxine