A smoke alarm has one big flaw.
It’s loud when you singe your morning Pop Tart - and it’s just as loud when there’s a raging fire in your kitchen.
The noise around Artificial Intelligence is similar, except that in place of a plastic disk on the ceiling we have some of the world’s biggest minds yelling at each other, arguing about what is on fire.
The AI alert is no longer limited to job fears; the smartest minds in the room are now stating, without doubt, that AI has the potential to destroy humanity .
Jacob Coxon, formerly a researcher at AI giants Anthropic and OpenAI, claims the people developing these systems believe that they could bring about our end by the close of the decade.
Evan Hubinger, another AI guru, says there is a greater than 10% likelihood of human extinction within the next decade.
Billionaire SpaceX boss Elon Musk and OpenAI co-founder Sam Altman say similar.
If your plumber arrived to install a new boiler and casually mentioned there was a 10% chance it would destroy civilisation, you would likely probe a little further before letting him reach for his tools.
So, before we hand AI the spanner, perhaps we should do the same.
Over the last 50 years, we have been led to believe that dangerous AI resembles HAL, from Stanley Kubrick’s 2001: A Space Odyssey refusing to open the pod bay doors.
Or Skynet, from the Terminator movies , launching nuclear missiles.
While this has delivered us some great cinema, it doesn’t help us understand a technology that does not need to hate us, nor require consciousness, in order to harm us.
The truly hazardous manifestation of AI won’t resemble The Terminator.
It’ll be a dull, invisible, efficient system buying, selling, diagnosing, hiring and firing at....
