AI hacking tools have got cheap, fast and easy to use. The organizations least able to defend themselves are now the easiest targets. According to The Verge AI, community hospitals, credit unions, small nonprofits and local retailers face a growing gap between what attackers can do and what defenders can afford. The most powerful AI security tools stay locked behind invite-only access for Big Tech.
A Small Nonprofit Takes the Hit
The Verge AI opens with Janice Malone, who runs Vivian’s Door, an Alabama nonprofit that supports underserved and minority-owned businesses. In March, people around the world started calling her. They’d received emails from her organization “begging for money,” and she hadn’t sent any of them.
Her third-party IT team took the systems offline for three days to find and close the hole. The bill came to about $3,000. She also worried she’d exposed financial data from the businesses she was trying to help. She still doesn’t know whether a human or an AI system was behind the attack.
“Who knows about the next vulnerability? You only know about the one that you’ve been hit with,” Malone said. “How do you protect yourself? I mean, really?”
Why the Math Changed
For years, one fact quietly protected small organizations: there were only so many skilled hackers, and they went after the biggest payouts. AI agents have broken that limit.
Here’s what has shifted, based on The Verge AI’s reporting:
- Scale: AI agents are now consistently strong at coding and security work, and they can run at enormous scale.
- Low skill needed: Attackers with limited AI knowledge can do “vibe-hacking” with automated systems.
- Shotgun targeting: Hackers who once went only after high-value targets can now hit everyone at once.
- One person, team-level output: In August 2025, Anthropic disclosed that a cybercrime ring used Claude Code to extort healthcare groups, emergency services, religious institutions and government entities, all within one month.
Anthropic threat intelligence lead Jacob Klein put it bluntly at the time: work that once needed “a team of sophisticated actors” can now be done by “a single individual” with agentic systems.
Michael Kleinman, head of US policy at the Future of Life Institute, summed up the change: “The limiting factor used to be that there’s a finite number of malicious hackers in the world, and that’s now no longer the case.”
The Defense Gap
What stands out here is the imbalance. AI is supposed to help defenders too. Anthropic’s Mythos is reportedly flagging so many vulnerabilities that Microsoft is struggling to patch them fast enough. But the top labs limit access to their strongest security models, including Mythos and OpenAI’s Astra, to a short list of players like Nvidia, Google, Apple, essential infrastructure providers and maintainers of critical open-source software.
Even with wider access, the price would likely shut out smaller organizations. So attackers get cheap, widely available models, while the best defensive tools go to companies that already have big security budgets.
Apollo Research CEO Marius Hobbhahn told The Verge he expects the real damage to land far from Silicon Valley: “A single person somewhere in a basement with one of the open-source models probably could hack a hospital and demand ransom… I expect the harm to be felt by a random Idaho hospital.”
Kleinman listed who’s exposed: community banks, savings and loans, credit unions, local hospital networks and local power grids. These are small institutions that deliver vital services.
What Smaller Organizations Should Do Now
You can’t match a Fortune 500 security budget. You can make yourself a harder target than the organization next door. Given how easy AI makes mass attacks, a few basics go a long way:
- Lock down email. Spoofed and hijacked email accounts, like the one that hit Vivian’s Door, are the cheapest attack to automate. Turn on multi-factor authentication everywhere and set up SPF, DKIM and DMARC records.
- Patch fast. AI is speeding up how quickly vulnerabilities get found, so exposure windows are shrinking. Automate updates where you can.
- Know your vendors. As Craig Smith of The Cool Hardware Company pointed out, small businesses rely on Microsoft tools and on outside procurement and delivery systems. Their weak spots become yours.
- Plan for downtime. Three offline days cost Malone money and business. Keep offline backups and a written response plan.
- Budget for incidents. A $3,000 surprise hurts. A ransomware demand hurts far more. Look into cyber insurance and managed security services built for small teams.
What Comes Next
The pressure is on AI labs and policymakers to close this gap, whether through cheaper defensive tools, wider access programs or shared threat intelligence for smaller institutions. Until that happens, local organizations should assume they’re on the target list. The full story, including more small-business perspectives, is at The Verge AI.