Well, this is EMBARRASSING!

An acclaimed author just wrote and published a well-reviewed book about truth.

Except…

It was accidentally full of LIES!

I spotted the story last week in the online edition of The New York Times.

It revealed that Steven Rosenbaum’s The Future of Truth: How AI Reshapes Reality – a nonfiction book contained misattributed or invented quotes.

One was by Kara Swisher, a technology journalist, who (apparently) said:

“The most sophisticated A.I. language model is like a mirror. It reflects our own morality back at us, polished and articulate, but ultimately empty behind the surface.”

However, she told the New York Times that she “never said that,” adding, “I also sound like I have a stick up my butt, according to ChatGPT.”

Another chapter contained quotes from How Emotions Are Made: The Secret Life of the Brain, by Lisa Feldman Barrett, a professor of psychology.

But Barrett has complained the quotes “don’t appear in the book and they are also wrong.”

Now, here’s the interesting thing…

Steven Rosenbaum has freely admitted that he used AI to help him write this book.

He’s not ashamed of that, or trying to hide it.

The problem is that he (or his editor) didn’t pay due diligence when it came to checking what AI generated.

“I used A.I. tools ChatGPT and Claude during the research, writing and editing process,” he says. “That does not excuse these errors, of which I take full responsibility.”

And that’s the big problem with AI.

THE CREDIBILITY GAP

The issue is not that people are using AI…

Because a LOT of us are!

Some estimates even suggest that over 1.3 billion people have accessed an AI tool in some way or another, which is about 16.3% of the global population!

And no wonder, when you think of the benefits…

It can save time… boost productivity….

And allow people without skills or education to access creative and business endeavours that would have been off limits a few years ago.

The problem is one of TRUST.

We now find it hard to trust anything we read, even when it’s a bestselling book by a credible author.

And if we don’t trust what we’re reading, we disconnect from not only the message in content…

But the person behind the content.

In the context of an online business, that’s obviously very important.

You might use AI to help you with planning, research, writing and design…

And that’s absolutely fine and I would encourage you to do it!

However, your subscribers, customers and website visitors need to KNOW there’s a real human behind your content…

Someone who is genuinely interested in helping them.

Otherwise,

HOW CAN THEY TRUST YOU?

They need to feel assured that you are giving them reliable information…

That you’ll deliver something unique that comes from your particular personality, experiences and insights….

And that you’re putting in work to make sure they have what they need to reach their goals or solve their problems.

What they DON’T want to think is that you’re simply entering a prompt and pasting whatever AI throws back at you.

Or that you’re an automated bot!

This is why we’re seeing such a big trend in ‘AUTHENTICITY’ right now.

Internet users are increasingly rejecting fake influencers, slick videos, and clever marketing hype….

And embracing real, everyday people – with all their flaws!

In an online world dominated by AI words and images, this is becoming such a pressing issue that there’s a new app called Sinceerly, which deliberately adds typos and grammatical errors into content.

The aim is to make your emails, social media posts and adverts sound more human, and less like they were written by ChatGPT.

Of course, we need to strike a balance…

I’m not sure that a deliberately typo-infested text would exactly look professional…

So it’s really about you adding in your personality, insights and observations.

Somehow you need to SHOW that you are a real human being who is actively working for the interests of your reader, viewer or customer.

That means being involved in the editing process – and not just accepting wholesale what Chat GPT or Claude gives you.

First, try and weed out the obvious AI-isms:

55 TELLTALE SIGNS YOU’RE USING AI

Here are some of them…

Common AI Phrases

• “It’s not about X. It’s about Y.” or…. “It’s less about X and more about Y.” (eg, “Not a tool, but a partner. Not a feature, but a philosophy.”)
• “In today’s rapidly evolving landscape”
• “Let’s explore” / “Let’s unpack”
• “At the end of the day…”
• “Grab a cuppa”
• “It’s worth noting…”
• “Notably…”
• “There’s a lot to unpack here”
• “Complex and multifaceted”
• “Intricate interplay”
• “Played a crucial role”
• “A testament to”
• “Cannot be overstated”
• “A plethora of” / “A multitude of”
• “From inception to execution”
• “ever-evolving”

Overused Words

• Delve
• dive
• foster
• underscore
• leverage
• harness
• optimise
• streamline
• empower
• facilitate
• amplify
• galvanize
• augment
• Pivotal
• crucial
• key
• tapestry
• pradigm shift
• ecosystem
• framework
• bandwidth

Punctuation & Emojis

• The em dash
• Colons (“The reason is simple: AI loves structure.”)
• The white tick in a green box (✅)
• Overuse of bold text
• Emojis on ✨ Bullet Points 🚀 Everywhere!

Tone & Voice

• Excessive positivity and agreeableness
• Lack of genuine opinion
• Neutral point of view
• Consistent emotional register without changes in tone or mood
• Absence of personal experience
• Lack of uniqueness or specificity

Structural Tells

• Paragraphs that begin with transition words like “Firstly,” “Moreover,” “Furthermore,” “Additionally,” and “Finally.”
• Paragraphs that are consistent in size.
• Similar sentence lengths and clause structures.
• The same idea repeated with different words – eg. “This is important. It’s crucial. It’s essential.”
• Overly diplomatic hedging, eg “It could be argued,” “Some might say,” “In many ways”.
• Performative balance where AI presents “both sides” of a question with such evenness that it avoids committing to a clear personal position.

If you use AI a lot you will begin to notice a lot of these recurring.

In my experience Chat GPT has the most obvious AI style and is the easiest to spot.

Claude is more subtle – though this can make it hard to tell when you’re relying too much on AI.

Whatever tool you use, here’s how I suggest you go about the editing process in order to humanise it.

BRIEF IT ON YOUR VOICE

Before you start a project, give your AI tool three or four examples of your own writing.

These could include emails, blog posts, social media updates… or whatever represents how you naturally communicate to your friends and family.

Then ask it to analyse your style and stick to it.

This won’t get perfect results because AI finds it hard to replicate a written style precisely, but it gives it a better starting point.

Also tell it what NOT to do…

Paste in a list of telltale words and phrases (like the ones I’ve given you above) and tell it to “Never use these.”

And say “Don’t use em dashes,” or “Don’t start paragraphs with transition words” to iron out some of the obvious stylistic ticks.

You should also feed the AI your opinions in relation to your business or the subject matter of the content.

For example:

• “I believe that small, sustainable farms produce better food than industrial ones.”
• “I think calorie counting is outdated and people should focus on nutrition.”
• “I think tech-bro culture is toxic.”
• “I think the obsession with open-plan living has gone too far.”
• “I believe the gym industry profits from making people feel guilty.”
• “I think businesses waste money on flashy websites when what they actually need is clear copy”
• “I believe children need to learn how to be bored.”
• “I think that Airbnb is ruining cities.”
• “I believe that exams only measure memory power, not true intelligence.”

Then instruct the AI like this: “Don’t hedge when it comes to expressing my opinion. Take a clear position.”

You’ll be amazed at the difference this makes when AI produces your first draft.

RUN YOUR DRAFT THROUGH A SECOND AI

Once you’ve got a draft from one AI tool (like Chat GPT), paste it into a different one (like Claude) and ask it:

“Does this sound like it was written by AI? If so, highlight the parts that give it away.”

Claude is often quite good at spotting ChatGPT’s habits, and vice versa.

But sometimes old-fashioned editing methods are the best way to counteract AI. So you can also try this…

READ IT ALOUD

Simply print out your draft and read it aloud!

You could even record yourself reading it on your smartphone and then listen back.

The key is to flag up anything that doesn’t sound like your natural speaking voice and tone… any words that you wouldn’t use in real life… and any phrases that sound robotic or stilted.

Or here’s another trick…

THE ‘TALK IT’ METHOD

Get AI to turn the draft into a set of notes – ie. key points laid out in an ordered list.

Print it out and open a voice recorder on your phone.

Now use the list of notes to talk through the topic in your own words and phrasing, as if you’re explaining it to a friend.

You can then transcribe that version and use that as the basis for your content.

Tell the AI, “This is a transcription of me speaking, maintain as much of this original voice in the piece as possible, including any verbal idiosyncrasies and flaws, but tidy it up for flow and structure.”

Now you should have a result that sounds much more like you, warts ‘n’ all!

ADD SPECIFICS

Finally, go through your drafted content and think of as many ways you can insert specific details that are unique to you.

• A personal anecdote
• A joke or comic aside
• A reference to something you watched, heard or read recently
• A mention of a family member or pet
• An allusion to a favourite film, author or piece of music
• An admission of a personal weakness
• An experience that a customer or client has had
• A result you got from your efforts
• An amount of money you made
• A trial or test result

These are what will make readers feel like they’re being talked to by a real individual.

And regardless of whether you use AI or not, they will ensure that any piece of online content is much more engaging and fun to read.