One of the most interesting points I have seen from Google recently is the idea of providing a unique point of view.
The basic message is this: do not just recycle what others have already said online, or what could easily be produced by a generative AI model.
That sounds simple enough, but I think it is worth thinking about properly.
Because in an AI search world, this point becomes more important than ever.
The Problem With Saying the Same Thing as Everyone Else
For almost any topic, there is already an established set of views.
If you search a topic online, you will often find the same points repeated across multiple articles, forums, videos and now AI Overviews.
The internet has already created a kind of default answer.
AI models have absorbed huge amounts of that existing material. So if we write an article that simply restates the same opinions, the same examples and the same conclusions, what are we really adding?
Probably not very much.
That does not mean the article is badly written.
It might be clear. It might be well structured. It might even be accurate.
But it may still be commodity content.
It is information that could have been produced by almost anyone.
And increasingly, it is information that could be produced by AI in seconds.
That is the real challenge for content creators now.
It is not enough to ask:
“Is this article correct?”
We also need to ask:
“Does this article add anything?”
Using AI to Find the Existing Consensus
This is where AI can actually be very useful.
Instead of using AI simply to write the article, we can use it to understand what has already been said.
For example, take a topic that has opinion, debate or judgement involved.
Something like:
“Is it fair that students in England have to pay for university education?”
There are already many familiar arguments around that topic.
Some people argue that students benefit personally, so they should contribute to the cost.
Others argue that society benefits from an educated population, so the state should fund more of it.
Some focus on student debt.
Some focus on fairness between generations.
Some focus on social mobility.
Some focus on whether university is still good value.
An AI model can summarise those existing arguments very quickly.
Google’s AI Overview may also show a version of the standard answer.
But that is only the starting point.
The useful question is not:
“What does everyone already say?”
The useful question is:
“What is missing?”
The Consensus Gap Method
I think there is a practical method here, which I would call the Consensus Gap Method.
It works like this:
First, take a topic.
Second, look at the existing views.
Third, ask AI what the common arguments are.
Fourth, look at Google, AI Overviews, Reddit, forums, videos or existing articles.
Fifth, ask yourself:
“What is everyone saying?”
Then ask the more important question:
“What are they not saying?”
That is where the opportunity is.
The gap may not be a completely original discovery.
It may not be something no human has ever thought before.
But it might be an angle that is underdeveloped, badly explained, too theoretical, too generic or missing the practical human reality.
That is where your article can become useful.
AI Can Suggest Angles, But You Still Need to Own the View
There is an interesting complication here.
If AI can suggest a fresh angle, does that mean the angle is not really fresh?
Not necessarily.
AI models do not only copy exact arguments they have seen before. They can combine ideas, spot tensions, suggest comparisons and create new framings from existing material.
So AI might say something like:
“The real issue is not just whether students should pay for university. It is whether we are treating university as a public good, a private investment, a route to social mobility, a workforce training system or a rite of passage.”
That may be a useful angle.
But the article does not become yours just because AI suggested that sentence.
It becomes yours when you think about it and decide what you actually believe.
Do you agree with that framing?
Can you improve it?
Can you challenge it?
Can you connect it to real examples?
Can you explain why it matters?
Can you add something from your own experience, judgement or observation?
That is the difference between asking AI for an opinion and using AI to help develop your opinion.
The Best Use of AI Is Often Conversation
This is why I think the best results from AI often come from a back-and-forth conversation.
A simple prompt can produce a simple article.
But a proper conversation can develop an idea.
You start with a rough thought.
The AI responds.
You agree with some of it.
You reject some of it.
You add an example.
The AI helps structure it.
You push back.
The idea becomes clearer.
You realise what you actually think.
That process is very different from saying:
“Write me an article about this topic.”
In that case, AI is doing most of the thinking.
But in a conversation, AI becomes more like a thinking partner.
It helps you shape the idea, but the direction still comes from you.
That distinction matters.
The Cake Mix Metaphor
This brings me back to a metaphor I have been thinking about.
A lot of people say that if you use AI to write content, you should add your own voice afterwards.
That advice sounds reasonable.
But I think it misses something important.
If AI creates the outline, writes the article and decides the main points, then adding your own voice afterwards is a bit like adding icing to a cake.
The cake itself is still mostly AI.
Your contribution is decoration.
You might improve the flavour slightly.
You might make it look more personal.
But the main ingredients were not yours.
My preferred approach is different.
I want my own thinking to be in the cake mix from the beginning.
That means using AI before the article is written.
Not just after.
It means having the discussion first.
Exploring the issue.
Testing the argument.
Adding my own examples.
Bringing in my own observations.
Developing my own view.
Then, when the article is finally drafted, the AI is not simply producing a generic article.
It is turning a human-led conversation into a structured piece of writing.
The cake mix already contains my ingredients.
What Does “Adding Your Own Voice” Really Mean?
This also raises another important point.
People often talk about adding your own voice as if voice is mainly about writing style.
They might mean:
- make it sound more casual
- add a personal sentence
- change a few phrases
- make it less robotic
- add a story at the beginning
- rewrite some paragraphs in your own words
There is nothing wrong with that.
But I think voice is deeper than sentence style.
Your voice is also:
- what you notice
- what you care about
- what you question
- what examples you choose
- what you think is missing
- what you disagree with
- what conclusion you reach
So if those things have already come from you during the conversation, then your voice is already in the article.
Even if ChatGPT writes the final draft more fluently than you could.
That is an important point.
Because, honestly, ChatGPT can often write better than I can in terms of structure, flow and clarity.
It can organise thoughts quickly.
It can make the argument cleaner.
It can remove repetition.
It can turn a messy discussion into something readable.
So if I have already provided the thinking, the experience, the viewpoint and the direction, I am not sure it always helps to then force the article back into my rough natural writing style.
In some cases, that might make it worse.
Sometimes Adding Your Voice Afterwards Can Weaken the Article
This may sound slightly controversial, but I think it is true.
If the article is already based on your real thinking, adding your own voice afterwards may not always improve it.
It can make the article more repetitive.
It can make the argument less clear.
It can add unnecessary personal comments.
It can make the writing feel forced.
It can turn a clean article into a rambling one.
That does not mean you should publish AI output without review.
You still need to read it carefully.
You still need to check that it says what you mean.
You still need to remove anything that does not feel right.
You still need to make sure the final article reflects your actual view.
But the goal should not be to make the article sound exactly like your unedited writing.
The goal should be to make sure the article reflects your thinking.
That is the difference.
A Better Question to Ask
Instead of asking:
“Have I added enough of my own voice?”
I think the better question is:
“Was my own thinking involved deeply enough before the article was written?”
And also:
“Could someone else have produced this same article from a simple prompt?”
If the answer is yes, the article may still be too generic.
But if the answer is no, because the article contains your examples, your argument, your metaphor, your doubts, your experience and your conclusion, then it is much stronger.
It does not have to be badly written to be human.
It does not have to contain awkward phrasing to be authentic.
It does not have to preserve every rough edge of your original thought process.
It just has to be built from something real.
This Is How AI Can Help Create Non-Commodity Content
The irony is that AI can either increase commodity content or help us escape it.
Used badly, AI can produce endless average articles that repeat the existing internet consensus.
Used well, AI can help us see that consensus more clearly.
It can show us the standard views.
It can reveal the familiar arguments.
It can help us spot what is missing.
It can challenge our thinking.
It can help us develop a stronger point of view.
But the human still has to bring something to the table.
That might be experience.
It might be judgement.
It might be practical insight.
It might be a personal story.
It might be a clearer metaphor.
It might be a more honest conclusion.
It might simply be the ability to say:
“I think the usual answer misses the real issue.”
That is where useful content begins.
Final Thought
For me, the best way to use AI in content creation is not to treat it as a vending machine.
It is not just:
“Enter keyword, receive article.”
The better approach is conversation.
Use AI to explore the topic.
Use it to understand the existing views.
Use it to test your own thinking.
Use it to find the gaps.
Use it to structure your ideas.
Then let it help you write clearly.
That way, the final article may be polished by AI, but it is not just AI content.
Your thinking is already inside it.
Your viewpoint is already inside it.
Your judgement is already inside it.
Your voice is not just added as icing at the end.
It is baked into the cake from the beginning.