The rise of generative AI has sparked a lot of debate about its potential impact on writing, but the influence it may have on reading has remained largely undiscussed.
The more I’ve thought about this, the more I’ve come to believe that adding this second piece to the equation reveals a much more profound implication of generative AI technology:
AI won’t just change text production; it’ll fundamentally transform the way that people exchange ideas, information, and stories. It’ll do this by redefining the relationship between content, writer, and reader.
My hope is that this transition will dramatically lower the barriers to sharing and digesting great thinking between different people and communities.
The issue with current text-based communication is that concepts and words are inherently bound to each other. This forces the writer to pick an imagined target audience — and by definition, this means they’ll be writing in a format that proves unideal for many of the people who would like to engage with their thoughts. Regardless of who’s reading or what their goals are, the writer’s words remain locked in their original length, style, vocabulary, and language.
The consequence of this is that readers must make all kinds of awkward adjustments to the way they process things in order to accommodate the fact that the text is fixed. The most common of these behaviours is probably the exhausting practice of skimming. You might even be doing it right now.
The crux of the transformation I’m predicting is that the future of reading and writing will be dominated by AI-powered mediation between the person who generates concepts and the person who consumes them. Instead of adjusting the way we read to accommodate fixed texts, our texts will adjust themselves to accommodate the way we wish to read. If we want to read fast, read with added context, or read without industry jargon, the text will simply rewrite itself accordingly.
It seems to me that we’re likely to see the transformation of the relationship between reader and writer begin on the consumption side. In the coming months or years, I believe new features will emerge in our digital reading tools like browsers and e-readers etc. that will allow us to specify the way we’d like to engage with the content we’re taking in. For a while, writers will hold onto concepts like authorship and definitive final drafts, but sooner or later it’ll become clear that notions like these are destined for obsolescence. The only exception will be work like poetry for which the authentic and original words are the whole point. Which texts fall into that bracket however will be determined at the reader’s discretion.
As generative AI gets better and better, I’m willing to bet that we’ll soon also be able to decouple ideas from their original medium. Even at the time of writing in 2023, AI can already illustrate existing texts with reasonable accuracy provided it has a little human guidance. Perhaps in the future, it will be possible for ideas, information, and stories to rearrange themselves seamlessly from text to video to VR and back.
For this reason, I think we’re likely to see the words “writer” and “reader” largely replaced by something closer to “idea originator” and “idea consumer”. The idea originator will be responsible for the core thinking and intent, while the idea consumer will be responsible for engagement in a format of their choosing. The exact words or pictures etc. will be handled by an AI middle-man.
Earlier I alluded to ways that generative AI may change text-based content consumption, and I’d like to unpack that in a little more depth. I think the following six changes will define the near future of text-based idea sharing. Picture the first three as buttons on a PDF reader:
Change #1: Read quick
We’re going to replace the tiering and mistake-prone practice of skimming text with a “Read Quick” button that will automatically remove extraneous material. If we choose, we’ll be able to read anything from any source at a relaxed pace, trusting that every word on the page is genuinely worth paying attention to. Perhaps this button could also include a field where we could explain our specific goal for reading the text so that the condensed adaptation would be suited to our aims. It would be like Control+F on steroids.
Change #2: Read deep
In addition to the “Read Quick” button, short texts like blog posts will also come equipped with a “Read Deep” button, which will deliberately extrapolate the original text into something that takes longer to read. It’ll do this by synthesizing additional angles on the original ideas. As the technology gets really good, it may even be able to add further conceptual depth and quality of analysis.
Change #3: Read accessible
This new set of buttons will allow the reader to adjust the difficulty of vocabulary and the complexity of sentence structure. This could make academic or jargon-filled writing far more accessible to people who work in other disciplines. It could also make texts from earlier periods in our language’s development much easier to grasp.
Change #4: Seamless translation
As these AI-based software tools perfect translation, language barriers in text consumption will disappear. All texts will automatically load in your language of choice. It won’t be a sloppy translation like the kind we currently have the option to turn on within a Chrome browser tab. It’ll contain all the nuance and emotion that would be provided by the best human translators today — and it’ll apply to everything instantly. As consequence, text-based communication with people who speak different languages will be completely seamless.
Change #5: Automatic updating
Texts will never again be out of date. When AI tools gain access to the internet, books and articles will be able to automatically update all the facts and figures they contain. This will give content about data-focused subjects like climate science a much longer shelf life.
Change #6: Adaptation to personal reading history
With the assumption that all of the above will make paper-based reading obsolete, the door will open to create text that contextually adapts itself based on our unified digital reading history. This will enable things we read to assist us in drawing meaningful connections with things we’ve read in the past. Imagine what it would be like to read a biography of Steve Jobs that’s tuned with the knowledge that you’ve already read Ed Catmull’s book about Pixar. The result would be worth more than the sum of its parts.
Would the changes I’m proposing bring about any difficulties? Of course! To start with, we’d have to find a way to prevent this from supercharging filter bubbles. We’d also want to think about how this might infringe on an author’s right to control the way their ideas are presented.
I’m confident that these challenges can be overcome as long as we take a thoughtful approach to execution. I’m also certain that the reward for doing so would be well worth the effort. By integrating AI tools into the way we share and consume ideas, a new dynamic of idea exchange will emerge that’ll dramatically increase the value our civilization gets from the time we invest in media-based communication. My hope is that this transition will also significantly lower barriers to people of all stripes who wish to engage with each other’s best thoughts. Just imagine the possibilities for shared understanding that this might create across cultures, generations, and areas of specialty. It’s going to change the world.
Designable.blog 