AI-assisted hacking puts small hospitals and businesses at risk

In March, callers warned Janice Malone that emails asking for money were coming from her nonprofit, Vivian’s Door. The Alabama organization took its systems offline for three days while its IT provider investigated and fixed a vulnerability. The work cost about $3,000. Malone still does not know whether AI played any role in the attack.

Her uncertainty reflects a wider concern: AI tools are helping people carry out cyberattacks faster and at greater scale. A person with limited technical skill can use automated systems to help probe for weaknesses, while experienced hackers can go after more organizations at once. In August 2025, Anthropic said a cybercrime group used its Claude Code tool in attacks on organizations including healthcare providers, emergency services, religious institutions, and government entities.

Small organizations have fewer resources

Large companies may be able to hire dedicated security teams and use advanced tools to find vulnerabilities. But some of the most powerful cybersecurity models, including Anthropic’s Mythos and OpenAI’s Astra, are available only to a limited group of organizations. The tools may also be too expensive for smaller institutions even if access expands.

That leaves local hospitals, banks, retailers, and nonprofits facing the same kinds of threats with smaller budgets and fewer IT staff. Vivian’s Door had to pay for emergency help while suspending its work. At the Maryland grocery co-op Takoma Park Silver Spring Co-op, hackers have used the online checkout system to test stolen credit cards, generating fees even when transactions are declined.

Healthcare faces especially high stakes

A cyberattack can affect patient care as well as data. Hospitals rely on digital systems for medical histories, allergies, medication lists, and other information; an outage can delay treatment and force nearby facilities to take on more patients. Fisher-Titus Medical Center in Ohio has one cybersecurity analyst, hired two years ago, according to its operations and information chief, Linda Stevenson.

AI does not create every vulnerability, and the source does not establish that it was involved in the attack on Vivian’s Door. But if it makes attacks easier to carry out, smaller organizations may have less time and fewer resources to keep up.