The headlines were impossible to miss:

Subscribe now for unlimited access.

Login or signup to continue reading

"Rogue AI hacks Medicare".

The phrase is evocative. It conjures images of sentient machines operating beyond human control, making their own decisions and pursuing their own agendas.

It sounds like the opening chapter of a science fiction novel rather than a cyber security incident.

And that is precisely why we should be paying attention to the language being used.

Not because there are no risks associated with artificial intelligence. There are. Not because the Medicare incident should be dismissed. It shouldn't.

But because the way we talk about AI shapes the way we understand it, and ultimately determines where we assign responsibility when things go wrong.

The first thing we need to acknowledge is that AI does not "want" anything.

It has no desires, ambitions, emotions, fears, hopes or intentions. It does not wake up in the morning plotting its next target. It does not resent humans. It does not dream of world domination. It does not dream at all.

Yet increasingly, we hear AI described as though it were a person.

It "thinks"; "decides"; "attacks"; "wants".

This may seem like harmless shorthand, but it matters. Language is never neutral.

If an AI "wanted" to access Medicare, then responsibility appears to belong to the AI.

If an automated system pursued an objective in a way that its designers failed to anticipate, responsibility belongs to the humans who built it, deployed it, authorised it, monitored it and failed to secure the system it interacted with.

One explanation blames a machine.

The other demands accountability from people.

People are considerably less convenient to blame. Especially when they work within the "machine" of government.

Perhaps this tendency shouldn't surprise us. Human beings have always anthropomorphised things they struggle to understand.

We gave human characteristics to gods, nations, markets and nature. We spoke of Mother........

© The Examiner