I’ve been in the ‘info publishing’ business a long time…

I first walked through the doors of Agora/Fleet Street Publications for a job interview back in 1993… and it genuinely changed my life.

(Not least because the first person I met was my future wife… Heloise!)

And I couldn’t recommend a better business model…

You get to solve problems for people… help them achieve their goals… and build an ongoing relationship with them.

In return, you can generate a recurring income from multiple products – with huge scope for leverage.

The more courses, books, newsletters and reports you create, the more you can potentially earn.

And the beauty of it is….

You don’t need to do all the work yourself.

In the old days, you could assemble products from public domain material… or more likely have to commission experts to research and write content for you.

These days, it’s a LOT easier.

Because in the AI era, you can get intelligent bots to do the research, writing and design for you.

The resulting product appears on-screen in a matter of hours – and it can be sold digitally without printing, stocking or delivering a single item.

Which means pretty much anyone can now create and grow a profitable home publishing business from home.

However, there’s a major problem…

The barriers to entry might have come tumbling down…

But it also means that the floodgates have opened…

The internet is now drowning in AI-generated slop, created by lazy people who don’t really care about the content, or the real-life outcomes of the customers who read it.

To give you an example…

When AI Slop Gets Dangerous

At the end of last year, the AI-checking tool Originality.ai scanned 558 titles in Amazon’s herbal remedies section and found 82% of the books were probably written by AI.

Sue Sprung, a medical herbalist in Liverpool, told The Guardian:

“There’s a huge amount of herbal research out there right now that’s absolutely rubbish. AI won’t know how to sift through all the dross, all the rubbish, that’s of absolutely no consequence. It would lead people astray.”

This wouldn’t be particularly worrying if the books were about, say, bird watching or decluttering wardrobes.

But this is health advice!

Just imagine what would happen if an unchecked piece of AI content recommended something poisonous to a reader!

It could be only a matter of time before disaster strikes!

To give you an example, last year, a guy named Richard Wordsworth wanted to find out more about his adult ADHD diagnosis.

He was shown a link to a book on Amazon that was all about the topic.

But he was surprised to see that it contained blatant historical inaccuracies.

Not only that… the advice was quite shocking.

The book told him that his family would never forgive him for the emotional damage he’d caused… and that his condition meant that he was “four times more likely to die significantly earlier”.

Again, this is mental health advice!

It could prove disastrous if someone’s unverified, unchecked, poorly researched AI creation resulted in the harm of someone who trusted that book.

The whole POINT of information products is that you buy it because you expect the creator to have ‘sifted through the dross’ for you.

So this is what I want to help you with today.

If you have any intentions to create or publish information products using AI, I want to show you:-

• How to avoid the big AI ethics problem – and ensure that your products actually help people, not endanger them or mislead them
• How to make your product stands out from the tide of slop – so it’s an appealing and engaging product that’s unique, human and relatable.

First, the golden rule.

‘Think Like a Publisher’

When you put out content through a newsletter, e-book or course you DON’T need to be an expert in the subject matter.

You don’t need experience or qualifications.

And you don’t need to be a professional writer.

What you’re doing is gathering, collating and presenting useful information to a niche audience who need help.

That’s all absolutely fine. No problem with that.

But you need to understand that you are a publisher, which means that you are – ultimately – responsible for what goes out into the world.

The buck stops with you.

So you need to carry out due diligence and ensure that the facts are correct… the advice is sound… and the tips can be actioned in the real world.

Not only that…

If you want to sell enough products to actually make money, you need to ensure that you’re delivering genuine value to your customers.

That means showing them that there’s a real human agent at work, who cares about helping them, and who has put some thought and effort into it.

None of this is to say you shouldn’t use AI.

Quite the opposite…

You SHOULD use AI… but as a writing and research tool that you combine with your own input.

The result will not only be an ethical product that delivers what it promises…

But also it will be the foundation of a profitable, long-term business where customers trust you, review you positively and share links to your products on social media.

Let’s take a look at ways you can do this.

Research, Checking and Editing

Start by treating every factual claim in your AI draft as ‘unverified until proven otherwise.’

In practice it means running a quick search on anything specific including:

• Statistics
• Named studies and books
• Dosages/measurements
• Historical dates
• Legal requirements
• Quotes attributed to real people.

You can use Google for this, simply typing in the claim or statistic and seeing what websites come up to verify it.

Or you can go to an AI tool like Chat GPT, Claude or Perplexity, paste in the content and use this prompt:

“List every factual claim in the text above and tell me the source for each one. Include a link so I can verify.”

The AI will often admit it can’t verify several of them, which you should consider as ‘red flags’.

If a major claim can’t be traced back to a primary source (a peer-reviewed paper, a government body, a recognised institution), cut it or rewrite it as a general observation rather than a stated fact.

That said, not every claim needs to come from a primary source.

For quotes, historical context, reporting on events, or background explanations, a credible secondary source is okay – e.g. an established publication such as the FT, Guardian, Reuters, or a well-regarded specialist website.

You could also use this prompt: “What are the most common misconceptions about this topic? Does anything in my draft repeat myths or misleading advice?”

When it comes to making big claims, I find that Perplexity is usually very good at assessing them and modifying them.

Just paste a large chunk of text, then ask it to go through and assess the claims.

Here’s a long prompt you can use:

“I want you to review the following information (text pasted below). Please check:

Whether the claims made are factual or misleading — tell me what you think needs evidence or seems exaggerated.

Whether the advice is safe, reasonable, and backed by credible logic or research.

Any red flags in tone, persuasion tactics, or marketing language that could mislead buyers.”

If you find conflicting reports or advice, don’t treat it as a dead end.

Just take some time to compare the quality of the sources.

(This is where you are adding value to your AI-powered information product!)

In most cases, favour the most credible, recent, and methodologically sound source.

A useful rule is to ask:

• Which source is closest to the original evidence?
• Which one is more reputable and transparent?
• Is one newer, better designed, or more specific to the claim?
• Are they actually contradicting each other, or just talking about different contexts/sources/timeframes?

If you are publishing online, I recommend you list any studies and sources, with links to the originals, at the bottom of the text.

This actually makes your content more appealing to the Google algorithm, so could potentially help you generate more traffic.

Subjects Where You Must Be Careful

When you’re selecting content to publish, it’s important to understand the risks.

The following subjects demand a higher level of verified factual evidence.

• Health, medicine and herbal remedies
• Mental health, psychology and therapy
• Legal and financial advice
• Supplements, dosages and drug interactions
• Advice directed at children or vulnerable adults
• Anything involving significant physical risk (exercise programmes for people with health conditions, for example)

In these areas, AI is genuinely dangerous if left unchecked.

Either avoid them, or if you need to include this information somewhere in your product, make sure you go back to primary medical, financial and legal sources.

Always add a clear disclaimer that the content is for informational purposes only and not a substitute for professional advice.

Humanising Your AI Content

It’s pretty easy to spot anything that’s been purely created by AI, with zero human input.

The content lacks a distinct personality… the voice is flat or cliched… and there are classic AI ‘ticks’ that people see repeated over and over.

For example…

• Repetitive sentence structures. For example, the constant use of this format: ‘It’s not X, it’s Y.’
• Generic claims with no real examples.
• Lists that feel padded out.
• Advice that lacks practical steps or specifics.

But the biggest overall giveaway is content that comes with no personally, offering advice that any reader could find with themselves with a quick Google search.

This is where you need to edit and enhance your AI draft with the kind of details that only YOU could provide.

• Anecdotes from your own life
• Insights you’ve gleaned from experiences you’ve had
• References to books, films, art, music or hobbies that you enjoy
• Your own results and outcomes from the advice
• Case studies or examples you’ve found yourself
• Humour and witty asides that you’ve personally added in
• Specifics from your life like your location, age, family status or job role.
• Your personal opinion on a topic
• Mentions of your worldview or moral outlook
• What you personally think is the most important takeaway from the advice
• Admission of mistakes you’ve made, or flaws in your personality

It’s okay if you lack writing confidence because you can always write these out in basic English, then get AI to tidy up the sentences for you.

Just make sure it doesn’t then turn it into AI-speak.

You can use this prompt:

“Tidy up the grammar and sentence flow in the passage below, but do not change the tone, and keep any informal language exactly as I’ve expressed it.”

Once your content is written, edited and humanised, it could be worth running it through Originality.ai, a tool for AI content detection and plagiarism.

It’s available as pay-as-you-go for $30, or on a Pro plan at $12.95 a month.

A Checklist for Non-Fiction Content

Before you publish anything, I recommend that you run through these questions:

✓ Does every factual claim trace back to a credible source?
✓ Is the advice actionable in the real world?
✓ Have you added at least one real human example or story?
✓ Does your voice appear consistently throughout the text?
✓ Would you be comfortable if a knowledgeable professional in this field read it?
✓ Is there a clear disclaimer where professional advice might be needed?

There is, of course, a LOT more I can tell you about publishing information.

So it would be useful to know…

What is it that you are most interested in publishing – eBooks, newsletters, digital courses, membership sites?

And what would you most like help with when it comes to research, content-creation, and the publication process?

Let me know…