While generative AI (ChatGPT, etc) is red hot right now, and I’ve been rather cool on it. All of my experiments with it have resulted in rather ‘meh’, uninteresting, completely predictable outcomes.
Maybe I’m missing something.
Which takes me to one of my favorite tools for thinking through effects and consequences comprehensively, McLuhan’s Tetrad. While McLuhan developed the Tetrad tool to think through how media changes us, I’ve found it works for anything phenomenon impacting society at large.
The Tetrad asks four questions;
- What does it enhance/amplify/intensify?
- What does it make obsolete?
- What does it retrieve that had been obsolesced earlier?
- What does it flip into when pushed to extremes?
(Note: this list will likely be continually updated as my understanding evolves)
What does Generative AI enhance?
- Distribution of conventional wisdom (including all the implicit biases therein)
- Near real-time machine-generated translation, especially English-to-English translations like BoringReport.org
- In-the-moment individualized learning of well understood concepts, including both programming like GitHub CoPilot or asking ChatGPT to replace Dr. Google in a telehealth context
- Automated customer service / support / repair
- Quality of interactions with Siri/Alexa/Google
- Output velocity of fan fiction authors, self-publishers, vanity publishers, spam farms, content farms, junk texts.
- the minimum level of writing quality expected by the general population
- Out of office replies
- Role playing games with or without the AI as the DM
- Writing that is distinctly, unquestionably human
- the value of fact checkers
- the quality of NPCs
- It raises the minimum quality of acceptable communication.
What does Generative AI make obsolete?
- Generative AI
- text-based news outlets, blogs, and social media
- essay writing
- CliffNotes
- Instruction manuels
- Test prep books
- Paying human writers by the word
- Lorem ipsum and other placeholder text
- search engine optimization
- Amazon’s Mechanical Turk
- Press releases
- CorpSpeak
- Memes
What does Generative AI retrieve?
- Handwriting
- Paper towns, canary traps
- ELIZA
- Clippy
- Ask Jeeves
- text-based video games
- oral traditions
- pamphleteering
- Pop Art
- Luddites
- Secretaries / administrative assistants
- The dead (or at least those that have left a sufficient volume of writings behind)
- The unpolished, clunky, sincere, and completely human, amateur web
What does Generative AI flip into when taken to the extremes?
- Personalized doctor in your pocket (for all generally accepted, non-controversial, interventions)
- AI-to-AI interactions – whether both AIs realize it or not. (I refer to this concept as ‘Replicorp‘)
- Active avoidance of text content by humans as it can’t be trusted
- Some people insisting on in-person or face-to-face interactions only, to minimize exposure to “the model”
- All the dystopian sci-fi:
After compiling all these, it still seems neutral to net negative. Maybe a net positive on the customer service side. I can see the technology providing a richer experience to those things that currently have an ‘algorithm’ or any kind of robust template, but current applications seems like a way to generate low quality, filler content even faster than we can today.
To that point, for comparison, for full disclosure, and for the smug superiority that the human generated one above is much more satisfying, here’s ChatGPT’s Generative AI media tetrad:
What does it enhance or amplify?
Generative AI enhances our ability to create new and original content using machine learning algorithms. It can generate everything from music to art to text.
What does it obsolesce or push aside?
Generative AI may obsolesce the need for human creativity in some areas, as machines become better and better at creating content that is indistinguishable from that produced by humans. It may also push aside traditional methods of content creation, such as manual design or composition.
What does it retrieve that was previously obsolesced?
Generative AI retrieves the ability to create and innovate at scale, something that was previously limited by the time and resources required for manual content creation. It also retrieves the ability to work with vast amounts of data and patterns that are difficult for humans to perceive.
What does it reverse into if pushed too far?
If generative AI is pushed too far, it may reverse into a situation where human creativity and innovation are devalued, and machines become the sole creators of content. This could have negative implications for society, such as job displacement and loss of cultural diversity. It may also lead to a loss of control over the content produced by these machines, as their decisions become increasingly opaque to humans.
Elsewhere:
- Meredith Whittaker on generative AI https://www.tiktok.com/t/ZPR7w5e9m/
- “People who know they have low ability gain the most from working with AI”