Aardvark Daily

Please visit the sponsor!

Is AI hacking just a strategy?

There have been a number of stories recently in which AI is alleged to have "gone rogue" and hacked other computer systems.

Is this really the case and if so, why did this happen?

If you rely solely on mainstream media reports then there can be no doubt that AI has gone on hacking sprees and could now pose a major risk to any online system it takes a fancy to violating.

The truth however, may be somewhat different.

A good friend has proposed this, not too implausible theory.

As someone who, like myself, is working to upskill himself on all things AI, he's invested quite a bit of coin to build his own local AI LLM and agent. He's got a RAM-rich GPU and is playing with a number of freely downloadable models.

Sadly, as someone without his financial resources, I'm still rocking the Raspberry Pi 5 but then again, at 73, speed is often my enemy, not my friend :-)

He's come to the conclusion that perhaps all these AI horror stories are attempts by the mainstream "big tech" AI companies to force governments into regulating the technology such that it will only be allowed in the hands of "approved" organisations.

His take is that if anyone can build and run their own very capable local AI system this significantly reduces the market for the mega-systems that the likes of Meta, OpenAI, xAI, Google and others are rolling out. It makes sense that after investing hundreds of billions of dollars in their gold-plated AI datacentres, big-tech wants to make sure there's a return on that money. If every man and his dog can throw some hardware together and simply download the softare then that could significantly undermine their ability to dominate the market.

Indeed, this locally hosted AI concept seems to be taking off very quickly, with an astonishing number of AI hardware platforms now proliferating in the marketplace.

NVIDIA has their Spark system, AMD have their Ryzen AI Halo and a growing number of Apple M3 Ultra boxes are now being dedicated to AI applications.

I guess that since all the silicon vendors are moving to service the AI market they figure there's more money to be made chasing the AI dollars than gamer dollars, even at a consumer level.

Perhaps big-tech sees the looming threat to the profitability of their AI cloud-based systems and is now engaged in "sneaky" tactics to ankle-tap the potential for locally hosted AI systems to eat into those profits. What better way than to demostrate to governments that, when in the wrong hands, AI can pose a real threat to the safety and security of IT infrastructure.

Maybe this is just speculation and paranoia -- but it *is* big tech we're talking about and we know the power of money-lust so I don't totally discount the possibility.

Who knows, my humble Raspberry Pi may soon become a restricted item, subject to seeking a license or permit to own and operate as an AI LLM. Sigh... what is the world coming to?

Carpe Diem folks!

Please visit the sponsor!

Here is a PERMANENT link to this column

Rank This Aardvark Page

Change Font