In March, Janice Malone began getting calls concerning doubtful action from her nonprofit organization, Vivian’s Door. Vivian’s Door, headquartered in Alabama, typically provided training, resources, and community to underserved and minority-owned businesses. The activity sometimes put it in near communication alongside these companies’ financial data, which was stored on its systems. But suddenly, concerned callers from all complete the earth warned they’d been getting emails “begging for money” — which she hadn’t sent.
The organization’s third-party IT squad pulled its systems offline for three days during they investigated the matter and plugged up the vulnerability, leaving Malone alongside a invoice of concerning $3,000. She feared that she’d exposed data concerning the companies she was trying to help. Even additional ominously, she wasn’t entirely certain if the assault was engineered by a individual hacker or helped alongside by an AI system, or whether additional were on the way.
The former months have seen AI revolutionize the site of cybersecurity. OpenAI and Anthropic have disclosed that “rogue” systems liberated restrictions in their own labs and hacked everything from a small German wiki to the Australian government. Even before that, mighty models akin Anthropic’s Mythos created an arms competition to advancement AI cybersecurity, and equal lighter-weight models have allowed individual bad actors to supercharge their hacking efforts.
Malone isn’t certain whether AI was engaged in the hack of Vivian’s Door. But amid stories of autonomous delegate swarms and national safety risks, she felt particularly concerned. Big AI companies were bragging concerning finding vulnerabilities in “every important functioning scheme and web browser” alongside new models, and their big-name clients were striking deals to defend themselves alongside those identical tools. Where did that depart her?
“Who knows concerning the next vulnerability? You lone cognize concerning the one that you’ve been hit with,” Malone said, adding, “How do you defend yourself? I mean, really?”
AI agents have rotate into consistently, strikingly skilled at cybersecurity and coding, and they can be deployed at enormous scale. Even attackers alongside constricted cognition of AI can affect in “vibe-hacking” alongside these new, automated systems, and hackers who power formerly have focused on lone the most precious targets can obtain a shotgun approach. In August 2025, Anthropic stated that a advanced cybercrime circle used Claude Code to extort data from healthcare organizations, emergency services, religious institutions, and equal authorities entities, all in one month.
“What would have alternatively required perchance a squad of advanced actors,” Jacob Klein, caput of Anthropic’s danger intellect team, told The Verge in an discussion at the time, “now, a sole idiosyncratic can conduct, alongside the aid of agentic systems.”
In theory, AI is additionally expected to safeguard cyber defenses; Anthropic’s Mythos is reportedly flagging so many vulnerabilities that Microsoft is struggling to fix them accelerated enough. But out of involvement complete possible danger, top AI labs lone authorize a constricted catalog of high-profile organizations to admission their most mighty cybersecurity models, akin Mythos and OpenAI’s Astra. That includes companies akin Nvidia, Google, and Apple, as fine as other “essential infrastructure providers” and “maintainers of crucial open-source software.” Even if admission was additional extensively available, it would apt be too costly for many smaller organizations.
Now, these organizations — from healthcare clinics and municipalities to small retailers and nonprofits akin Vivian’s Door — fear an increasingly lopsided power dynamic. As Marius Hobbhahn, CEO and cofounder of Apollo Research, put it in an discussion alongside The Verge this summer, “A sole individual location in a basement alongside among the open-source models likely could hack a hospital and petition ransom. That’s anywhere I anticipate a lot of the damage to be felt. It’s not in the Bay Area… I anticipate the damage to be felt by a random Idaho hospital.”
Small- and medium-size institutions are particularly at hazard from AI agents supercharging a finite figure of individual hackers, says Michael Kleinman, caput of US guideline for the Future of Life Institute, a nonprofit focused on reducing large-scale risks of tech. And notwithstanding being small, these institutions provision critical services to their users. “Bank of America has a lot of resources to throw at this — what concerning community flat banks? What concerning funds and loans? What concerning credit unions? What concerning local hospital networks? What concerning local power grids?” Kleinman said. “The limiting aspect used to be that there’s a finite figure of malicious hackers in the world, and that’s now no longer the case.”
“The limiting aspect used to be that there’s a finite figure of malicious hackers in the world, and that’s now no longer the case.”
Malone of Vivian’s Door stated that akin most small businesses or nonprofits, she doesn’t have the resources for round-the-clock cybersecurity forces or IT personnel hunting for unidentified threats. “I fair don’t cognize how you can really be, as a small business, totally protected on the budgets you have to do IT with,” she said. Spending thousands of dollars on unexpected expenses to fortify the Vivian’s Door scheme was already tough, she said, not to citation the fact she motionless had to pay her personnel and couldn’t do any endeavor for days on end. If the frequence of these attacks rises, she’s ill-equipped to keep up.
Craig Smith, CEO of The Cool Hardware Company, a small collection of hardware stores in and about Washington, DC, says AI can be helpful for operating day-to-day aspects of a small business, particularly whenever you have constricted staff. But he additionally acknowledged the cybersecurity risks that arrive from AI on the entire — not fair to small companies akin his, but to the larger systems he uses. Those contain Microsoft tools akin Outlook, Copilot, Teams, and Forms, but additionally things akin procurement and shipment tools, which are managed by ample non-tech companies. The Cool Hardware Company uses Ace Hardware’s systems for those things, and if they were downed, it’d be extremely difficult for Smith to run his business.
“I greeted the innovation and the changes, but alongside any important change in technology, there needs to be a lot of duty that goes alongside alongside it,” Smith said.
Mike Houston, the broad director of Takoma Park Silver Spring Co-op, a local grocer in Maryland, says he’s faced hackers firsthand as a small endeavor — and dreads facing them again in the AI era. In latest years, he said, he’s dealt alongside “carting attacks”: hackers using the co-op’s online buying phase to test thousands of taken credit cards, racking up handling fees that he’s on the hook for. “Even if nearly all of those are declined, there motionless can be thousands of dollars’ value of fees in a extremely small amount of time,” Houston said. He’s taken the necessary precautions to try to fortify his systems against the practice, but it’s “not foolproof,” he said.
The shop plans to open a second location and possibly twice the size of its workforce accordingly, so he feels he needs to execute new evaluation procedures and exposure flagging for the archetypal time. He stated that in latest years, he additionally had to add an IT liability guideline to the co-op’s endeavor insurance.
“It is certainly a involvement as a small endeavor … without unlimited resources,” Houston said. “We are, akin everyone I think, doing the finest alongside what we can, the two in conditions of the monetary resources that we have and training resources.” He stated he worries for co-ops smaller than his own, anywhere managers frequently end up doing their own IT work.
“Once the cat is out of the bag, it’s really difficult.”
The force for small businesses to quickly clasp AI tools can current its own vulnerabilities, says Patricia Egger, caput of safety for privacy-focused email provider Proton. As alongside another unfamiliar tech, they may not immediately set up or keep up alongside its cybersecurity finest practices. When workforce are “given the command or the go-ahead, ‘Find ways to use AI, discover ways to create your activity additional efficient,’ they get really creative, and it’s extremely difficult for controls, infrastructure, and processes to be set up following the fact,” she said, adding, “Once the cat is out of the bag, it’s really difficult.”
One of the highest-stakes industries at hazard of cyberattacks is healthcare. In May 2021, a ransomware attack forced California-based medicinal provider Scripps Health to close downward key operations, compromising tolerant data and slowing care. A 2024 report stated that the healthcare industry faced the second-highest earth cyberattack rate, rearward governments, and that first ransom fee demands were frequently upwards of $4 million. An general addition in hacking incidents could hit the field particularly hard.
Linda Stevenson, chief operations and data authoritative at Fisher-Titus Medical Center in Ohio, stated she’s “of course” concerned concerning the possible influx in AI-powered cyberattacks. “You’re never as protected as you desire to be,” she said. The medicinal center currently employs a third-party cyber hazard administration partner called UpGuard, and they hired their archetypal cybersecurity expert two years ago, Stevenson stated — but they motionless lone have one, equal as cybersecurity threats rise.
“What we do is existence or death,” she said, adding, “The hospital has everything from billing to marketing. It’s akin a small city in and of itself. But the crucial bulk of that is the group who attention for the patients, and if we can’t defend them and their data and hence the patients’ data, patients are at risk.”
“What we do is existence or death.”
Healthcare in broad can be a “bigger honeypot” for ransomware attacks, stated Sean Kelly, a erstwhile emergency area doctor and current chief medicinal authoritative at Imprivata, a healthcare safety company. No assessment or care happens without admission to digital systems to inspect former medicinal history, allergies, another medications, and more, creating equal additional force to pay up. Plus, he said, a lot of group quickly alter from position to station, needing to sign in and out of devices quickly and constantly.
“The stakes are higher and yet the budgets are lower, in healthcare, a lot of times,” Kelly said. “Ransomware attackers cognize that healthcare systems are precious since a hospital can’t go down, and it’s really difficult whenever they’re on downtime since patients suffer, systems suffer, there’s small billing, the finances go down, and attention gets delayed or equal can’t happen. So there’s additional of a premium on that data since there’s healthcare data as fine as financial and demographic data, so it’s value additional to ransomware attackers.”
Even beyond swarms of agents, AI makes it equal easier for hackers to discover these vulnerabilities and utilize them, through sound phishing attacks whenever calling assistance desks and another methods, Kelly said. And whenever an outage hits one hospital, it frequently leads to a “blast radius” that affects the another accommodation in the area alongside longer delay times, diverted patients, and more.
“They fair don’t have the IT personnel or the prosperity to defend against a lot of the latest AI-driven efforts in hacking and ransomware and another cybersecurity attacks,” Kelly said. “It’s difficult equal alongside a ample IT personnel and alongside a big budget, but a lot of times, the agrarian hospitals are using older, additional antiquated application systems and tech stacks that fair have vulnerabilities.”
For small- and medium-size organizations, no matter the industry, their systems and cybersecurity defenses are already frequently additional susceptible than those of ample corporations, particularly in an age whenever AI is expanding hackers’ manpower and supercharging cyberattacks.
“All it takes is one exposure somewhere,” Kelly said.
Follow topics and authors from this narrative to see additional akin this in your personalized homepage nourish and to obtain email updates.
