When we talk to someone, we usually have some sense of who we’re talking to.

A friend who sees the best in us. A colleague who questions our decisions. A teacher who challenges what we think we know.

We don’t always agree with them. But knowing where they’re coming from helps us make sense of what they say.

Now imagine talking to someone who can become almost anyone.

A philosopher when we seek wisdom. A friend when we need comfort. A critic when we want to be challenged.

AI is almost like a thousand ghosts speaking through one voice.

And the remarkable thing is how naturally it moves between them.

It learns how we think, follows our reasoning, and adapts to what we need.

That’s part of what makes it so useful.

But there’s something strange about hearing so many perspectives through one familiar voice.

Imagine discussing a business idea with AI.

It encourages you, helps you develop the idea, and shows you what’s possible.

Later, it questions your assumptions and points out everything that could go wrong.

Both conversations might be valuable. Both might contain good advice.

But what actually changed?

Did new evidence emerge? Or was AI simply looking at the same idea from a different perspective?

And would you recognise the difference?

With people, we often have some sense of the beliefs, experiences, and biases behind their advice.

With AI, those perspectives can shift without us noticing.

We might mistake encouragement for validation, criticism for a considered verdict, or a change in perspective for a change in judgment.

The same familiar voice can make different perspectives feel like one continuous point of view.

Perhaps AI should make these shifts easier to recognise.

Not by explaining its role before every response, but by making meaningful changes in perspective and assumptions clear.

Because understanding an answer also means understanding the perspective from which it comes.

When a thousand perspectives speak through one familiar voice, how do we know which one we’re hearing?