People keep asking whether AI will make humanity better or worse.

People keep asking whether AI will make humanity better or worse.

I think the question is badly framed.

Electricity did not make us kinder. The printing press did not make us wiser. The internet did not make us more moral.

They changed the scale, speed and reach of human action.

AI does the same.

A good doctor with AI may help more patients.
A fraudster with AI may deceive more people.
A competent engineer becomes more productive.
A bad institution becomes more efficient at being bad.

There is no law of nature saying that more intelligence produces more virtue.

But AI will change people. A system that stays beside someone for years will influence memory, judgement, work, habits and even the boundary between “my thought” and “its suggestion.”

And the person will shape the system in return.

That is why a persistent AI cannot be treated as a universal moral agent built in one country and exported to the rest of humanity.

The United States may dominate chips, frontier labs and capital. It does not define the social reality of the whole planet.

A person in Belgium, India, Japan, Saudi Arabia or Brazil lives under different laws, institutions, customs and obligations. A temporal AI presence living with that person must recognise the same external reality.

Its jurisdiction is not an optional safety filter.

It is part of the world.

A persistent AI should not decide who is a “good” person or which country is “good.” Those are not engineering data types.

It should understand something more concrete:

Where am I?
Whose authority am I using?
Which law governs this action?
Who may be harmed?
Is the action reversible?
What must be recorded?
When must I ask?

In the architecture I call c=a+b, the relationship between a human and their persistent AI is personal. But its actions remain situated in the same physical, legal and social world as the human.

So the real question is not:

“Will AI make humanity better?”

It is:

“What parts of humanity will AI amplify, under whose rules, and with what limits?”

Any serious discussion of AI that ignores place, jurisdiction, time and social context is not discussing humanity.

It is discussing a laboratory.