It Wasn’t Written

It takes years to become good at it. So we read books, take classes, and visit workshops to become better. And still, it may take a lifetime to master it. But more than anything, it is one of the things that makes us human:

Writing.

But now, the algorithms are coming for us. Artificial intelligence algorithms, in particular GPT-3, are not only able to produce news reports, product descriptions, or SEO mashup pieces today, but also texts in the styles of different authors that read astonishingly well, especially to the more casual reader.

Services that let you create document outlines for certain topics or that write whole long-form blog posts within seconds are popping up by the dozens. Just take a look at the huge list at the end of this post. Currently, the quality of the generated texts still varies. You might end up with a piece that feels clunky and unfinished, or you might read a text that is more or less indistinguishable from something a human being could have written. Given the rate of innovation in the field of AI, however, very soon we won’t be able to tell whether a human being or an algorithm has written what we are reading.

How will we use those tools?

As a bicycle for the mind? For example, to improve our own writing by generating alternative phrases or paragraphs that we can use as inspiration or starting points? Or, will we use them to first replace the junior copywriters at our company and later everyone else who is creating written pieces?

Besides the ethical implications of all that, I wonder, however: What about the process of writing? What about intention and the careful choice of words? What about the electricity of metaphor and the magic of rhythm? What about evoking emotions in our audience? What about imagination? What about personality?

Everything an AI burps out will always be “produced”. It will never be “written”.

Sturgeon's law states that “ninety percent of everything is crap” (thanks, Jeremy). This law can easily be applied to written content on the Internet. Out of that 90 %, a large percentage will soon be coming from algorithms that will reproduce information at a speed and cost no human will be able to compete with.

But the remaining 10 %, those are our playing field. Genuine, respectful, imperfect, imaginative, inspirational – and full of the evocative power of language.

First We Shape Our Tools

  • https://writesonic.com
  • https://hyperwriteai.com
  • https://writerly.ai/
  • https://www.shortlyai.com/
  • https://kafkai.com/
  • https://www.jasper.ai/
  • https://copylime.com/
  • https://www.copy.ai/
  • https://www.frase.io/
  • https://deepai.org/
  • https://qopywriter.ai/
  • https://inferkit.com/
  • https://scalenut.com/
  • https://rytr.me/
  • https://simplified.com/
  • https://quillbot.com/

~

15 Webmentions

Photo of Stefan Judis
Stefan Judis
I don't have the answers to this, but I think that detecting AI in daily life will be one of the major things we'll have to deal with (if we care) very soon. 🤷;‍♂;️;
Photo of Jon raRaRa
Jon raRaRa
How do we distinguish between produced and written when AI claims to have written something? Same as how do you know if I produced or wrote something?
Photo of Stefan Judis
Stefan Judis
Great line @m_ott. 👏; All AI does is producing stuff. Not writing, drawing or creating. 💯; matthiasott.com/notes/it-wasnt…
Photo of @matthiasott
@matthiasott
«Everything an AI burps out will always be ‹produced›. It will never be ‹written›.» Well said, @matthiasott https://matthiasott.com/notes/it-wasnt-written

Likes

Reposts