Could ChatGPT End Up *Raising* the Quality of Writing and Public Discourse?

Since the launch of ChatGPT, there’s been a lot of speculation on impact that it will have on writing. Many have expressed concerns that there could be a proliferation of low-quality content, as it’s now easier than ever to “write” an article or post using AI. But there’s another, often overlooked, possibility: What if ChatGPT actually elevates the quality of public discourse?

In this blog post, we’ll explore the idea that, just as past technological advancements have led to higher standards in various aspects of life, ChatGPT could raise the bar for written communication.

History Repeating Itself: the Printing Press and Public Education

Before diving into AI language models, let’s revisit the 15th-century printing press. Back then, many elites worried that printed materials and increased literacy would degrade public discourse. However, these concerns proved to be unfounded. The printing press facilitated knowledge democratisation and led to significant social, political, and cultural advancements.

Similarly, as public education spread and literacy rates increased, elites worried that mass literacy would lead to a decline in cultural and intellectual life. They were concerned that it could result in the spread of misinformation, superficial knowledge, or the “dumbing down” of culture. Sound familiar? However, much like the concerns about the printing press, these fears proved to be largely unfounded. Increased literacy rates and access to education have generally been seen as positive developments, contributing to social progress, economic growth, and the democratisation of knowledge.

Today, these same concerns are being voiced about AI-generated text. Let’s scrutinise those claims a bit closer and make sure they’re not just repeating the same old elitist claptrap.

AI Language Models as the New Washing Machines

Throughout history, advancements in technology have often led to changes in societal expectations and standards. But rather than lowering the bar, technological advancements have tended to raise it.

Think about the way household appliances like washing machines and dishwashers revolutionised cleanliness. Before their widespread adoption, maintaining a clean home was a time-consuming and labour-intensive task. But as these appliances became more accessible, society’s standards of cleanliness rose. People started expecting cleaner homes.

Like washing machines, could ChatGPT end up increasing our standards for writing hygiene?

Similarly, computers and the Internet made it much easier to store and retrieve information. This changed the way we approach education, moving the focus from rote memorisation to critical thinking and reasoning skills. It’s no longer enough to simply recall facts; we now expect students to analyse and synthesise information.

Lastly, consider the impact of word processors on writing and editing. Before, creating and editing written documents was a laborious, manual process. You couldn’t hit backspace to fix a typo on a typewriter, and spell-check tools weren’t available to identify errors. The advent of word processors revolutionised writing by allowing writers to easily edit, format, and revise their work. As a result, our expectations for error-free writing have risen.

We Already Deal with an Abundance of Low-Quality Content

In today’s digital world, it’s already incredibly easy to write and publish a blog post. There is a vast amount of content online, much of which is low quality. But the vast majority of posts never reach a large audience. We already use various different mechanisms to sort the wheat from the chaff. For example:

  • Search engine algorithms prioritise authoritative and relevant content in search results.
  • Community-driven platforms like Reddit use upvotes to surface valuable content.
  • Plain old word-of-mouth. People share and promote content they find interesting, informative, or entertaining. Most people share with their family and friends, while influencers and thought leaders share content they believe their audience will appreciate.

While it’s true that AI-generated text would put additional pressure on these systems, we should remember that we’re already dealing with this issue and have developed ways to manage it. As these tools and platforms continue to evolve, they will likely adapt to the changing landscape, ensuring that high-quality content remains discoverable and accessible.

“Why Didn’t You ChatGPT It?”

Instead of only worrying that ChatGPT will result in a flood of low-quality writing, we should also consider that ChatGPT might improve the overall quality of writing.

Imagine a world where not using ChatGPT (or a similar AI tool) to proofread and edit your work is akin to neglecting spell check. Maybe we’ll expect students to “ChatGPT” their assignments before handing them in. No one thinks using spell check is “cheating”, and perhaps we’ll eventually see the use of AI language models as best practice, rather than as an unfair shortcut.

It’s certainly true that ChatGPT makes it easier to generate a subpar post, but it correspondingly makes it easier to generate an excellent post. Our best thinkers are not necessarily the best communicators. I’m sure you know of geniuses who find it hard to express themselves in a way that ordinary people can understand. As a tool that can generate well-structured, comprehensible, and grammatically correct text with ease, ChatGPT could contribute to a higher standard of writing across the board.

A Balanced Perspective

So, to summarise:

  • The history of the printing press and public education shows how concerns about new technologies degrading discourse have arisen in the past โ€” and have proven to be unfounded.
  • We already have mechanisms to sort the wheat from the chaff in terms of content quality (though admittedly, AI language models could put a lot more pressure on these).
  • Technological advancements, like washing machines and word processors, have tended to raise societal expectations and standards rather than lower them.
  • ChatGPT could well become a standard tool that people use to improve their writing and communication.

I don’t mean to dismiss the risks that AI-generated text poses. But it’s important to take a balanced perspective and recognise that we’re already facing a lot of those risks. While it’s certainly plausible that ChatGPT may increase those pressures, let’s also recognise the possibility that ChatGPT might end up raising the bar by improving the quality of writing and public discourse.

How do you think AI will affect the quality of public discourse? Are you bullish or bearish?

If you enjoyed this post, you might also like:

4 thoughts on “Could ChatGPT End Up *Raising* the Quality of Writing and Public Discourse?

  1. Very well put together thoughts. I admire and agree with your forward-thinking approach in regards to ChatGPT. The technological luxuries of today will become the new baseline for the future. Humanity has always built upon itself in this regard, and ChatGPT is no exception.

    I am curious to here your thoughts on your relationship to LLMs in respect to this blog. Are you worried about ChatGPT’s capability to fulfill this blog’s primary value proposition of detailed summaries? Are you yourself leveraging ChatGPT in writing your blog posts?

  2. Thank you, Zak! Yes, I am worried about ChatGPT’s amazing summarising abilities, but not *too* worried as: (1) I don’t make money from this site and never had high hopes for doing so; and (2) I had always hoped to be able to deliver value in ways other than just providing summaries, anyway. But the summaries definitely helped me get visitors so if people stop searching for them, that will impact the site.

    And yes, I’ve been using ChatGPT to help with my writing – including with this very post.

    I’ll write up a separate post (or posts) going into more detail as I have mixed thoughts about it all and writing a post will likely help clarify my thinking ๐Ÿ™‚

  3. I also think it could likely increase it. Often when I’m thinking through a problem I want to talk to someone about it, but the issue could be quite objectively boring (lol) or technical, or just otherwise difficult. Prior to ChatGPT, I’ve either been lucky I’ve known someone who might be interested through work or something else, or just never really sat down and worked it through.

    But in the last two months I’ve probably “chatted” with ChatGPT about 15 times on these types of issues. For me at least, I’ve found that’s a great way to explore an issue and think about the different angles and where problems might be. For the smartest people, maybe it won’t be helpful for exploring their thoughts (but as you say, could be helpful for communication). But for the rest of us, I think it’s easy to underestimate being able to chat through an issue with an “intelligent”, “interested” partner.

  4. I suspect bouncing ideas off chatbot is helpful even for the smartest people. Perhaps even moreso because, like you, they may face a greater struggle to find others knowledgeable and interested enough to talk to about their esoteric or technical issues.

    What’s fascinating is that one of the earliest chatbots (ELIZA) worked simply by reflecting people’s statements back at them, with some probing questions to help the conversation along. Many people were convinced they were talking to a real human and claimed to have meaningful, therapeutic experiences with “her”. Anyway, my point is that to chat with ChatGPT, you have to clarify your thinking enough to put it into writing, and that process itself can be helpful. And ChatGPT can obviously add a lot more value than ELIZA.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.