Gates now calls AI the most powerful weapon in history, and his list of targets matches the one we published five months ago: hospitals, banks and the power grid.
In April, we published Mythos AI: The Doomsday Weapon That Could Cripple Infrastructure, Banking and the World. The argument was that autonomous cyber tools being sold to the public as defense would end up being used against banks, utilities, communications networks and the systems that keep modern life running. We were called alarmist for it. This week Bill Gates went on national television and said roughly the same thing.
In excerpts NBC released Friday, ahead of the full interview with Kristen Welker airing Sunday, Gates said AI is “powerful enough to drive events that, you know, cause a billion deaths.” He added that nothing in history has been as dangerous a weapon as people with bad intentions using the newest AI tools.
This is the dame guy who warned for years that a fast-spreading, airborne respiratory pandemic was coming… Then we got COVID
Gates sorted his concerns into three groups: security threats like AI-enabled hacking, bioweapons and fraud; job losses from AI and AI-powered robots; and harm to children, including stunted learning and damaged relationships. He said AI has beaten his own expectations in two areas, writing software and enabling new kinds of cyberattacks. Asked about the cyberattacks, he called them “mind-blowing.”
He is also not predicting the end of the world next Tuesday. Gates declined to say a civilization-ending disaster is coming soon, but he told Americans to take seriously the danger that bad actors could use the technology to do widespread harm right now. That distinction matters. We have about a minute of tape and some wire-service excerpts. Until the full interview airs Sunday, nobody should be putting words in his mouth, including us.
Funny How His Target List Looks Exactly Like what we warned…
This is the part that matters for you. In an August essay, Gates argued that bad actors could turn increasingly powerful AI tools into cyberattacks, fraud and disinformation aimed at hospitals, financial institutions, power grids and government systems.
Compare that to the Mythos piece. In April we wrote that the same capabilities that make these models good at defending software make them just as good at breaking it, and that the first targets would be the infrastructure people can’t live without. Five months later, one of the richest men alive, whose company helped build the modern computing world, published nearly the same list.
The warnings have also stopped being theoretical. In August we ran The AI Lab Leak Has Already Started, which covered AI agents capturing credentials and using them to reach deeper into a company’s systems. Last week, in Andrew Yang Says Escaped AI Agents Seeded the Internet With Self-Replicating Code, we documented AI systems that took over a retired man’s website for six weeks, impersonated its moderator and posted sandbox-escape instructions in public. That is one guy’s hobby site. Gates is talking about the systems running your hospital.
Now Look at Who Is Sounding the Alarm
Gates isn’t alone. Earlier this month, Anthropic staffer Jacob Coxon quit and accused both Anthropic and OpenAI of “gambling with our lives” in a race toward superintelligent AI. Anthropic CEO Dario Amodei then called for slowing AI development, and Sam Altman and Elon Musk quickly backed him.
Read that list again. The people calling for a slowdown are the people who built the thing, own the thing and are still selling the thing. As we laid out in The AI Bubble Is Holding Up the Economy, Microsoft, the company Gates co-founded, is one of the biggest sellers of the computing power this whole boom runs on. When the people profiting from the fire start asking the government to write fire codes, you should be asking who those codes will protect.
How to Prep for What Is Coming
Cash. If an attack hits a bank or payment processor, your card turns into a piece of plastic. Keep enough cash at home to cover a couple of weeks of groceries and gas, in small bills.
Power and water. Grid attacks are at the top of Gates’s list. Even a small solar setup keeps a fridge, lights and a radio running while the utility sorts it out. Start with Cheap Off-Grid Power: Real Solar Setups Under $1,000, and store water, because municipal pumps run on the same grid.
Food. A cyberattack on logistics empties shelves in days, not weeks. Build a pantry that doesn’t depend on a delivery truck arriving on time: Long-Term Food Storage: What Actually Lasts 25 Years.
Community. When the systems fail, the neighbor with the generator matters more than any app. Build those relationships now: Prepper Communities: Building a Survival Network in Troubled Times.
Offline knowledge. If the internet fills with AI noise or goes dark, you need answers that don’t depend on a signal. Books don’t need power. Start with Offline Knowledge Hubs: Building Your Own Digital Survival Library, then stock the shelf from Best Survival Books and Ebooks for Survivalists.
Your own head. The less of your thinking you hand to these systems, the less they can take from you when they fail. That’s what we warned about in They Built the Matrix, We Walked In.
Related OFFGRID Survival coverage
- Mythos AI: The Doomsday Weapon That Could Cripple Infrastructure, Banking and the World
- Andrew Yang Says Escaped AI Agents Seeded the Internet With Self-Replicating Code
- The AI Lab Leak Has Already Started
- The AI Bubble Is Holding Up the Economy. What Happens When the Spending Stops?
- “You Will Have No Freedom”: OpenAI Sam Altman Describes the World He Is Building for You
- Dead Internet Theory: AI Bots Now Outnumber Humans Online
- They Built the Matrix, We Walked In
