Facts vs Opinions: The secret to talking about AI

The best way to think about AI is ... 

Think of its output as an opinion.

And think of data as facts.

And be very careful not to confuse one with the other!

Opinions are subjective. 

Facts are not.

While looking at a bunch facts, not everyone comes to the same conclusion, not everyone ends up with the same opinion.

Multiple ‘experts’ can have varying ‘opinions’.

For example...

My height is 5’6”. That’s a fact. That is data.

But am I tall? Or short?

That is an opinion.

It depends. It is contextual.

It depends on who I am being compared with.

One AI model may classify me as tall, another as medium, and another as short.

Think of AI as an opinion generator.

And a really good AI can generate an opinion that has the sheen of an actual fact.

Similar to how some ‘experts’ have opinions that seem just right.

Most people follow opinions, not data.

And when opinions become ‘expert opinions’, then for some people they are as good as facts.

Opinions are insights & inferences, derived from experiences, created by noticing patterns, and melding them together through intuition and logic.

AI models try to replicate the same process.

Opinions can change, and so can AI output.

With experience, exposure, training and relearning.

Even experts have bias, and their opinions have bias, and we tend to accept that bias while listening to that expert.

We just need to be good at knowing which opinions matter.

Sometimes we learn this the hard way in real life.

Don’t confuse AI output with real-world data.

Most of the time it will nearly be, but don’t start with that assumption.

Always remind yourself,

That the AI output is an opinion.

If you want to act on it,

Should you treat it as an opinion?

Or as a fact?

Use your judgement, just don’t assume they are same.

Share with your peers if you think this helps.

If you need help navigate this space, then visit skolasolutions.com for additional tips, or to connect with me.