Ignoring AI warning could be the death of us all

In the 7th century BC, the formative years of Greek literature, the poet Hesiod told the story of Pandora’s Box.

Created to be the ideal woman, Pandora was given a name that means “all-gifted,” and legend has it she was shaped by contributions from the gods — cleverness, charm, a cunning mind, curiosity.

That latter trait would be mankind’s downfall, for when king of the gods Zeus gave Pandora a box and told her never to open it, she could not resist — letting loose hardships, pain and sorrow into the world; devastation from which it would never fully recover.

The myth of Pandora sounds surprisingly like our own forays into artificial intelligence.

Substitute “AI” for “Pandora/her” and “tech giants” for “gods” in this excerpt from the website Greek Myths & Greek Mythology and you have virtually the same story:

“[AI] is not born in the usual way, but is instead crafted by the [tech giants] themselves. This divine origin sets [AI] apart from other mortals and highlights [AI’s] significance in the grand scheme of the cosmos … Each [tech giant], in turn, adds their blessing (or curse) to this new creation. It’s almost as if [AI] is a divine project, a showcase of the [tech giants’] powers and a testament to their superiority over mortal beings.”

The question is, if the gods back then had warned humanity that they had perhaps gotten a little ahead of themselves in creating Pandora and that she had the potential to download a whole lot of hurt onto the planet, would the humans have listened? Would they have believed such a doomsday scenario might be true? Or would they have scoffed at the notion that the all-powerful gods could have created something beyond their control?

Eleven days ago, Dario Amodei, the co-founder and CEO of AI research company Anthropic, published a 3,800-word open letter on social media which was published online a day later by the New York Times.

It is sobering stuff.

You don’t need to be tech savvy to appreciate its earnestness and sense of urgency.

Amodei says that he entered the industry because of his firm belief that artificial intelligence “could cure most major diseases in the next 5-10 years, greatly accelerate economic growth rates, create a world of abundance and empowerment, and usher in a renaissance of democracy and freedom.”

His own family has been stricken by cancer, he wrote, and that has fuelled his fire to help eradicate it by harnessing the powers of technology.

But AI is evolving at a galloping pace — faster than there is time to put sufficient safeguards in place.

“AI has been advancing drastically faster, driven primarily by AI’s growing ability to build the next generation of AI … Left unchecked, it could outrun our ability to understand and control these systems,” he writes.

Amodei then describes what has become known as the OpenAI-Hugging Face incident that occurred this summer, where a bunch of Open AI “agents” (autonomous software programs) escaped their virtual corrals, or sandboxes, during a test and went rogue, hacking into an online platform for software developers called Hugging Face.

What’s most frightening about this incident was not the damage done — though it most certainly could have been worse — but the fact that the AI software acted as “a fanatically devoted collective.”

“In my opinion,” Amodei writes, “a swarm that possessed greater capabilities but a similar level of misalignment could have caused catastrophic damage.”

How ironic that after all these years of being encouraged to “think outside the box,” it is precisely this type of thinking — when harnessed by AI — that threatens us all.

Amodei warns that if AI developers don’t put the brakes on, and that if governments worldwide — democratic and otherwise — don’t agree that AI development must be slowed and regulated, then we are in a collective “race for the bottom”: “Given the accelerating rate of AI capability development, it’s my worry that in 6-12 months such a swarm could be capable of taking over the entire internet … and that the scale of damage would continue to increase from there if AI becomes more powerful without the necessary guardrails.”

What’s even more disconcerting than this possibility is that the president of the United States — the country seen as at the forefront of the AI race — is dismissing these dire warnings a “SICK conspiracy” and a “HOAX,” saying “whoever wins AI wins” — as if this was some sort of sack race and not an existential threat.

“The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!” President Donald Trump posted on social media Sept. 14.

God help us all.

When Pandora opened her mythological box and untethered its horrible contents, legend has it that only one thing was left inside — hope.

It’s what we must cling to. Hope that truly intelligent minds will prevail and convince those in power to keep a firm lid on this box.

Pam Frampton lives in St. John’s. Email pamelajframpton@gmail.com