Armageddon and artificial intelligence

Just when some online worry-warts and so-called “climate experts” are starting to ease a little on warnings that the end is nigh thanks to an era of global boiling, we have a new doomsday scenario to contend with which could make all that redundant by the end of the decade.

Maybe you missed it because of all the other bad news in the headlines. Perhaps you’ve been too busy worrying how you’ll meet your next mortgage repayment while your house value drops in the wake of Labor’s Federal Budget, or your business has gone belly-up. How will you pay your rising rent or your next electricity bill? But that’s all irrelevant if you blieve the latest claims that humans will soon be wiped out.

Not by climate change or a new virus more deadly than Covid, but by Artificial Intelligence! Yes, that AI bot that helps you with information and friendly advice on a huge range of topics within the blink of an eye could be part of a gathering silent army in burgeoning data centres with the ultimate aim of making us mere dumb humans obsolete.

That is, if you believe the recent high-profile warnings from prominent AI safety researchers claiming there is a significant risk that Artificial Intelligence could soon lead to human extinction.

These stark predictions, primarily driven by researchers at top frontier labs like Anthropic, OpenAI, and Google DeepMind, have shifted the debate from whether an existential threat is possible to how quickly it could occur.

The warnings escalated following a wave of high-profile industry resignations and public statements by insiders who claim AI labs are locked in a dangerous, unregulated race to build superintelligence.

They include Jacob Coxon, a former OpenAI & Anthropic Researcher who posted a viral warning stating that the professionals building AI "earnestly believe that it could kill us all by the end of the decade". He added that top executives publicly understate their fears but express intense panic privately.

Evan Hubinger, a leading Anthropic Alignment Scientist, publicly backed Coxon’s claims, stating he personally calculates a greater than 10 percent chance that unaligned AI will wipe out humanity within the next 10 years.

Alex Turner, a former Google DeepMind Scientist went further, estimating that the probability of an AI takeover of human civilisation is roughly "one-in-three".

Crikey, have these blokes been sipping too much Kool Aid or maybe watching some streaming shows such as The Capture on ABC iView which gives the message that we can’t believe what’s in front of our eyes especially on our TV screens, or the older Person of Interest  on Netflix with two warring AI machines battling to control our lives?

Good entertainment, possibly containing some predictive truths akin to Nostradamus, but even they don’t go so far as to suggest that AI is secretly planning to wipe us all out.

I’m no scientist or computer whiz, but even I can see a lack of logic to such wild claims. A couple of years back I wrote an article Beyond Mad Max with Dystopian AI which pointed out the huge power demands of ever-expanding AI data centres.