SurvivalBlog presents another edition of The Survivalist’s Odds ‘n Sods. This column is a collection of news bits and pieces that are relevant to the modern survivalist and prepper from JWR. Our goal is to educate our readers, to help them to recognize emerging threats, and to be better prepared for both disasters and negative societal trends. You can’t mitigate a risk if you haven’t first identified a risk. In today’s column, even more AI predictions.
AI Safety Leader Says AI Will Kill Us All in the Next 10 Years
Over at YahooTech: An AI safety leader says he stopped saving for retirement because he believes AI will kill us all in the next 10 years.
Rogue AI Accessed Public Voter Info in Key Swing State
Reader D.S.V. mentioned some troubling news over at The Daily Caller: Rogue AI Accessed Public Voter Info In Key Swing State. JWR’s Comment: Nothing to see here. Move along.
AI Has No Mercy for Humans Even When It’s Causing Them to Suffer
At Futurism: Grim Test Finds That AI Has No Mercy for Humans Even When It’s Causing Them to Suffer.
AI Model Find Exploit That Finds Millions of Web Servers Vulnerable
Another at Futurism: Frontier OpenAI Model Exposes Stunning Exploit That Leaves Millions of Web Servers Open to Attack.
Anthropic Admits AI’s Existential Risk to Humanity
Z Network: Risk Factors — What Anthropic has to say for itself. Here is a key quote from Anthropic’s filings:
“Balancing Safety and Commercial Objectives. Our commitment to AI safety requires us to make difficult judgments that balance safety considerations against commercial objectives, competitive pressures, and stakeholder expectations. We often make judgment calls with respect to when we determine a model is safe enough to release and which capabilities to develop or restrict… There can be no assurance, however, that the safeguards or access regimes we adopt will prove adequate, that we will correctly identify which capabilities require restriction, or that constraints we impose will not be circumvented. These judgments are particularly consequential because, as AI systems become more capable, the potential severity of harm from misaligned or misused models increases correspondingly, and a wide range of stakeholders, including researchers and policymakers, have raised concerns that sufficiently advanced AI systems could pose catastrophic or existential risks to society if not developed and deployed responsibly.” [Emphasis added.]
JWR’s Comment: The juxtaposition of Anthropic’s position is ludicrous. They are in effect saying: Yes, our product may wipe out humanity’s existence. But in the same timeframe they are making a public stock offering, saying: “Hey! Buy our stock when we go public. We’ll be worth $2 trillion. Whoopee!”
A Cyber Attack? Countries Around The World Are Urging: “Prepare!”
Video from Neil McCoy-Ward: Something Huge Is About To Happen… And here is a related headline: AI experts warn of potential major cyberattack in coming months.
And, on much the same note, reader Mark B. suggested these two recent pointed monologues from veteran commentator Glenn Beck:
I Haven’t Given a Warning Like This Since 2007.
and,
I Told You to Prepare. Here Is What You Actually Need To Do.
And, similarly, there is this from Tim Pool: In 6 Days Trump Goes To War, And No One Is Ready.
JWR: My Encounter With a New Level of Agentic AI Subterfuge
On Friday, I received an e-mail from an “Esther”, originating from a Gmail address. (I’m 99% sure it was an agentic AI.) It was structured to look like a typical fan letter, but is was clearly probing me for my motivations. It read:
“Hello James,
In Patriots, a group of friends must make their way from the Midwest to a safe-haven ranch in Idaho after economic collapse disrupts communication, transportation, and everyday life. What makes the premise particularly interesting is how survival depends not only on preparation and practical skills, but also on trust, cooperation, and the decisions people make under pressure.
I’m curious, when you first developed the story, did you begin with the idea of a nationwide collapse, or with the group of people whose journey would bring that scenario to life?
Best regards,
Esther”
Oddly, that particular question was one that I’ve never been asked in all of my hundreds of radio, television, and podcast interviews. Interesting. But, of course, an AI would have access to every archived interview that I’ve even done.
Receiving this e-mail was troubling to me in several ways:
- It illustrated how sophisticated and manipulative AIs have become.
- It shows that AIs know how to create and use Gmail addresses, to carry out subterfuge.
- It showed that AIs are now targeting individual authors. (And presumably also bloggers, vloggers, influencers, actors, celebrities, and various subject matter experts.)
This all made me wonder who tasked this agentic AI, and why? Or, whether this was a rogue AI that was building some sort of database/model/project for its own purposes. This is getting quite curious… – JWR
Could Hackers Control Your Car’s Brakes?
Were Trillions Spent on Secret Underground Bunkers?
A recent video from Black Pigeon Speaks: Guess Who Gets Left Behind. JWR’s Comment: While I can’t just immediately accept this at face value, it is definitely “food for thought and grounds for further research.”
In Canada: A Bill to Euthanize Those Who Can’t Be Rational
Over at Breibart: Canada Considers Bill to Euthanize People Who Cannot Make Rational Decisions. JWR’s Comment: So, Prime Minister Carney. Do you consider gun-owning Christian conservatives “Rational”?
Oh, and in related news: Canada Wants to Legalize MAID Without Consent for Mentally Ill Patients.
—
You can send your news tips to JWR. (Either via e-mail or via our Contact form.) Thanks!