Let's Talk About The Elephant In The Room - Writers and AI

Years ago, in another lifetime, when my children were small, I used to work for a US Federal Court reporter.  She lived in California and I lived in Australia.  Her job as a stenographer was to record every single word that was said (and by whom) for the court cases she was assigned to.

While sitting in the courtroom, she used a typewritery-kind of machine, but instead of letters on the keys, it had shorthand characters.  So when she typed … in shorthand … the machine would then automatically transcribe her shorthand into sentences (which were rarely perfect), then save that transcription onto a tape.

Every week, I’d get a packet in the mail from her.  Each packet had a couple of audio tapes containing the court case proceedings.

I’d take those tapes, put them into a tape player, then laboriously transcribe what I heard, sentence by sentence, into an Amiga (a super clunky, early version of a computer).  

I’d print out those documents and then post them back to the court reporter by mail.  The whole process used to take about 3-4 weeks roundtrip.

And then I’d wait to get paid.  And wait.  Then wait some more.  

Eventually I’d get a check in the mail for several hundred dollars, take it down to the bank and then wait even longer for the check to clear before I could access my money.

It was a fantastic job to have back then.  I could do the work while my kids were asleep.  I didn’t have to be in the same country as the court reporter in a time when remote working was practically unheard of.  And the pay was superb.  The biggest challenge was dealing with the tedium of hitting the back button until I could decipher what someone was saying, then edit the garbled version the machine had produced.

Little did I know that in just a few short years, this whole process could be completed in a matter of hours with the use of AI-based transcription tools.

That’s great, right?

Well, recently, I’ve been having a number of disturbing conversations with both writers and translators who report that their revenue is dropping and that it’s getting harder to find clients.  Invariably, they say that it’s because their (former) clients are using AI tools to do the work they used to do and at a fraction of the price.

Maybe you can relate?

If so, it’s completely understandable if you’re feeling like the sky is falling.  Or that it’s the beginning of the end for your writing (or translating) career.

I’ve heard enough people say they’ve been affected that I decided to look more deeply into how AI is impacting writers (and translators) these days.

And I came across a report from Andrew McAffee who was the 2023-2024 visiting fellow for technology and society at Google.

In his report on The Economic Impact of Generative AI, MAffee reports that:

"On one popular online freelancing platform, both monthly income and number of jobs for people offering 'writing-related services' declined in 2023 as online generative AI became widely available"​.

Cause for concern, right?

But there’s more to the report than that and much of it is positive.  

Basically, The Economic Impact of Generative AI report explores how generative AI, as a general-purpose technology, has the potential to really speed up economic growth. And that growth is generally achieved by increasing productivity across various sectors (including writers and translators).

How does generative AI get that increased productivity effect?  

Because AI technology can 

1. drastically reduce the time it takes to do certain tasks and 
2. create new job opportunities for those people willing to add working with AI to their existing skill set.

McAffee reports that:

"Professional writers, using a generative AI chatbot, completed occupation-specific writing tasks 37% faster with significantly higher quality"​

And also that:

"Translators, for example, often no longer begin their work by translating an entire text themselves. They instead start with a machine-generated translation, then audit, refine, and improve it."​

The Glass Half Empty View of AI

So if you want to go down the glass half empty scenario, McAffee’s research is saying that Generative AI can complete tasks faster than humans without a loss of quality, impacting up to 80% of jobs. 

Now that IS a scary number!

And it will definitely have an impact on a lot of people.  In fact, McAffee says that AI technology will have a quicker (and more profound) impact on the economy than innovations like the steam engine or electricity (both of which helped kick off the industrial revolution).

And like the Industrial Revolution, people who don’t upskill quickly enough will find that their existing skills are no longer needed.

But there’s also a positive side.  There are some benefits to be had in using AI as a writer.

The Glass Half Full View of AI

Here are just three of those benefits.

1. Productivity Gains - Getting More Done In Less Time

AI (and more specifically, Generative AI) has huge potential for helping us be more productive. 

Remember the quote from McAffe’s report?  That writers using generative AI complete tasks 37% faster and produce higher-quality outputs​?

Well, when you’re writing, “faster” can matter.  A lot.

In her book The Freelance Content Marketing Writer  (affiliate link), veteran writer Jennifer Goforth Gregory talks about the secret to bringing in big money from writing.  She says it’s not so much about how high a price you can charge, but rather how quickly you can complete a writing assignment.  

In other words it’s all about your internal hourly rate … how much you can earn per hour for each writing project.

A $1000 writing assignment that takes you 20 hours to research, write and edit is going to give you an hourly rate of $50 an hour.  Whereas if it only takes you 5 hours to get to the finished product, your hourly rate breaks down to $200 an hour.  

Plus you’ve got all that extra time you didn’t have to spend (if you were slower), which you can spend marketing yourself, sending out pitches or finishing off another assignment.

So anything that can help you get to a finished product more quickly is going to help you increase your hourly rate.  Which means you can either work less for the same amount of revenue or you can spend the same amount of time and bring in more money than before.

2. Giving Slower Writers A Boost

If you’re new(ish) to the paid writing game, it can be a struggle.  Not only is it harder and slower to write winning pitches, it can also take longer to get those pitches accepted and then written.  

But using Generative AI can help close the gap between you and someone who might be a more experienced writer, but they haven’t yet started using AI.  Not because it’s doing your writing for you, but because you’re using it in a way that retains your writing integrity AND saves you time.

3. Translators (and Some Writers) Already Use AI, But Could Do More

One of my collaboration project partners is veteran translator Jason Willis-Lee.  And he tells me that translators often begin doing a translation using translation software.  They then take that machine-generated translation, check it for errors, refine and improve it.

In some ways using that translation software is changing their role from being a full translator to being more of an editor or reviewer.

And it’s the same for writers.  If you do interviews as you prepare for a writing assignment, why in the world would you try to transcribe those interviews by hand!  

No.  You’d use a transcribing service or tool like Descript (affiliate link) or Otter.ai.

And is that any different from using a machine that translates shorthand into coherent English sentences the way my court reporter boss used to do?

Because here’s the thing.  

The value that I brought to the table in my work doing court transcriptions, wasn’t the fact that I could push a button or run a computer.  

No.

The court reporter told me (just about every time she talked to me) that what she really valued about my work, was my ability to listen to a garbled sentence and correctly recreate what had actually been said, so that the transcription was accurate.

And accuracy is everything when it comes to reporting on court proceedings.

Point of Clarity About Types of AI

I keep using the term “generative AI”, but there’s also predictive AI and they aren’t the same thing.

Here’s what ChatGPT said when I gave it this directive.   “Please state the difference, using non-academic language, between generative AI and predictive AI”.

This is as good as any definition I’ve come across in listening to podcasts and reading blog posts and reports, so I’m letting it stand as is.

Where I Stand On The Use of AI Tools

Personally, I find AI very useful.  I have a number of plates spinning in the air at any one time and anything that is going to help me get rid of dull, repetitive tasks is something I welcome.  

For instance, I use Descript (affiliate link), which I mentioned earlier.  Descript is a video and podcast editing tool and I use it to help with producing my Online Business Launchpad podcast episodes.  I use a very structured, step by step process to turn the video file into a transcript, shownotes, audio file and several video files for use on social media and my YouTube channel.

Does Descript do all the work?  No, I heavily edit what the AI produces.  But because it all starts with my voice and thoughts, I’m comfortable that even where I don’t rewrite stuff, I’m retaining my voice and “vibe” all the way through the finished product.

And Yet … We Could Be Right To Worry About AI

Despite the value I get out of using AI, I still have some concerns about AI overall.   And especially the impact it can have on us as writers or on our ability to earn a living writing.

Pale, Stale Information

My first major concern has to do with the information held by the large language models and how they are managed and updated (or not updated).

The Negative Spiral - A Decline in Content Quality

AI tools are built using large language models, which are essentially massive sets of data.  Unless that data is refreshed or updated on a regular basis, there’s this not-so-nice negative spiral of AI data going stagnant.  Until eventually what’s being said is so out of date or nonsensical or unoriginal that what it produces essentially becomes worthless … or even worse, causes harm.

Baked In Bias

Then there’s the reality that the people who put together the large language models have in-built biases that are reflected in the content that’s included and in how the algorithm responds.

I’ve heard too many stories about people getting really weird, skewed results that don’t reflect reality for many groups of people.  So it’s important to be mindful of potential biases and actively work to mitigate them when they pop up.

And what about errors in the datasets that haven’t been picked up?  Or information that’s outdated.  Or even misinformation that’s been deliberately included?  That probably sounds a little bit melodramatic, but it’s on us to check what we’re getting and use our brains to make sure we’re not swallowing everything whole that comes out of AI.  

I think that a little skepticism is not a bad thing here.

Remember AI Isn’t Human

And because it isn’t human it’s going to miss lots of little bits that we, as humans, instinctively pick up on.  Things like insights into human thought and psychology, emotions and cultural nuances.  I think that over time the AI will get better at this if, and only if, humans are aware enough to build those things into the large language models.

And need I add that as humans make new scientific discoveries, we’ll have to consistently update the AI.  Otherwise, critical information will rapidly get out of date.  That probably seems obvious, but do we know whether the processes and funding are in place to keep those AI tool updated?  

I’m not sure at this point.

AI Might Make Us Lazy

The second major concern I have is that AI is really appealing to the lazy side of our nature (or is it just me!).  It’s hard work to learn new skills.  And maybe we need money now, so we do anything we can to skip some of those learning steps.

Which means that if we start using AI as a crutch or a way to not have to learn something new or improve on what we already know how to do … we are going to be worse off.

Why?  

Because we won’t have those skills available to us if we’re ever in a situation where no AI is available.

But even worse, it could stop us from learning how to learn, which is an art unto itself.

So, unlike some people I’ve spoken with, I don’t believe that I’ll be replaced by AI any time soon.  However, I do think we need to be highly aware that there can be a dark downside to this amazing technology, which we’ve already touched on above.  Not to mention the still as yet unresolved legal issues about intellectual property rights and whether they are being stomped all over.

There is Great Potential for Good

But having said that, we’re already seeing that in some areas, the use of AI is a tremendous force for good.

I recently spoke with accessibility expert, Maxwell Ivey, about what his take on the use of AI is.  Maxwell (or Max to his friends) is blind.   Despite that, he actively consults and he’s also a paid writer.  And he mentioned something that I wasn’t even aware of.  

He said that Facebook has managed to do something that not many other software platforms have.  And that is to provide image descriptions that actually describe what’s in the image, rather than a keyword stuffed Alt text description people often use.

So if I was to wrap my thinking about AI into a nutshell, it would be this … AI is a tool.  We can use that tool in ways that over time will produce a negative outcome.  Or we can think things through for ourselves and come up with ways to use AI tools that are helpful, without losing our integrity.

Good or not so good outcomes.  The choice is yours.

So don’t be afraid to use AI, but always fact check.  Be authentic.  Keep writing and creating new content in your own wonderful, imaginative way.  And make time for more of your own writing by letting AI handle the repetitive, non-creative tasks

Because I know that for myself, the sheer joy of writing is something that I never want to lose.  Being able to use AI to give myself more writing time is a gift I’m delighted to accept.