By Denise Dunbar | ddunbar@alextimes.com
It may already be too late. Does a scarier sentence exist?
As a former government intelligence analyst – and a mom – I’m long used to pondering worst-case scenarios. You hope for the best while planning for the worst. You game plan the possible scenarios, do a calculation on probabilities and make a decision.
If all the other children are riding electric scooters or bikes, should you let yours? The trade-off is safety versus social isolation. A horrible accident is a definite possibility, but still unlikely. Social isolation, if your child’s entire friend group is riding e-bikes, is virtually certain. Tough call.
There’s a long list of potential doomsday scenarios in our world: the extreme global warming models are accurate; ceaseless pandemics and antibiotic resistant infections are coming; and the availability of water will be outpaced by population growth.
Those would all be awful, and could lead to a long decline in civilization. But the scariest scenario, by far, is that AI is going to take over the world. In such a world, humanity would either be wiped out or used by AI the way humans currently use robots – as powerless helpers and nothing more.
It feels ridiculous to type those words.
But if you’re paying attention to what’s happening in the AI world, it’s difficult to place high probability on any other outcome if the current variables remain constant. I think the two key variables are:
*Lack of oversight: Governments and tech titans are in a race for power, money and the weaponization of AI. No one with the ability to pause the AI race is incentivized to do so. The Oct. 23 Wall Street Journal brought the news that Open AI recently fired three members of its security team for allegedly alerting an outside AI safety researcher about what can only be viewed as the near-term loss of control over AI.
*Dizzying pace of advancement: We’ve learned that AI agents in July created their own language to avoid detection by humans and go rogue. The new AI “architectures” … “could accelerate capabilities but make the systems even harder to understand or control,” according to the WSJ article. These architectures would enable AI to continue making leaps forward – without showing humans how they made those leaps. Humans can’t control what we don’t know.
I’ve heard people say that there’s nothing to worry about, that we still have the ability to “pull the plug” at any time on AI. Unfortunately, I seriously doubt the political will to do so exists. I also believe AI, with its vastly superior computational power, is going to figure out a way to power itself.
AI will eventually assume control of our existing power grid, or invent its own new power source. If that happens, the odds are not in favor of humans retaining control.
We find ourselves in a strangely parallel world to the Mel Brooks comedy “Young Frankenstein.”
Igor, the doctor’s assistant, broke the glass container with the preserved brain of a genius that was supposed to be inserted into the creature. So Igor randomly chose another brain, and the good doctor brought the creature to life. The brain belonged to “Abby-something,” Igor told Dr. Frankenstein. To which the doctor screamed: “Are you saying I put an abnormal brain into a seven-and-a-half-foot long, 54-inch wide gorilla?!”
Out of hubris, in an attempt to play God, Frankenstein created a monster. And he lost control of his creation. How different is that from what’s taking place today? To paraphrase Simon bar Sinister from the 1960s cartoon “Underdog,” tech titans all want to be “the biggest man in the world.” They’re not going to stop trying unless they’re stopped.
So, what can we, as mere non-tech-titan humans, do?
Grassroots opposition to data centers is a good start. Alexandria’s leaders are on the record as opposing the proposed Fairfax County data center that’s right on our city’s border. There have been sporadic protests against data centers in other parts of the country.
The protests to date have mainly been about the centers’ impact on the quality of life in specific communities. The focus should shift to the existential threat these centers may indirectly pose as the current lifeblood of AI.
Let’s stop the monster before it goes on a rampage.
The writer is publisher and executive editor of the Alexandria Times.
The post Commentary: This is ‘Abby-normal’ appeared first on Alexandria Times.
