--- title: What Does Anthropic's AI Security Expansion Mean for Hospitals? description: Anthropic's AI now finds vulnerabilities at scale. Five Nines Technology Group explains what that means for hospital patching, risk, and security plans. image: https://blog.fivenines.com/hubfs/Blog%20banner%20Discovery%20Got%20Cheap@1x.png --- [Skip to the main content.](https://blog.fivenines.com/what-does-anthropics-ai-security-expansion-mean-for-hospitals#main-content) 1-855-817-5959    [help@fivenines.com](mailto:help@fivenines.com) [![FN\_Reverse\_Logo](https://blog.fivenines.com/hs-fs/hubfs/FN_Reverse_Logo.png?width=3022&height=849&name=FN_Reverse_Logo.png "FN_Reverse_Logo")](https://fivenines.com/) [![FiveNinesPrimaryLogo](https://blog.fivenines.com/hs-fs/hubfs/FiveNinesPrimaryLogo.png?width=3022&height=849&name=FiveNinesPrimaryLogo.png "FiveNinesPrimaryLogo")](https://fivenines.com/) - [Solutions](https://fivenines.com/solutions/) - [Managed IT Services](https://fivenines.com/solutions/managed-it-services/) - [Cybersecurity Services](https://fivenines.com/solutions/cybersecurity/) - [Cloud & Productivity](https://fivenines.com/solutions/cloud-productivity/) - [Projects & Enhancements](https://fivenines.com/solutions/it-projects/) - [Placement Services](https://fivenines.com/solutions/onsite-it/) - Industries - [Finance & Banking](https://fivenines.com/industries/banking/) - [Healthcare](https://fivenines.com/industries/healthcare/) - [Legal](https://fivenines.com/industries/legal/) - [Nonprofit](https://fivenines.com/industries/non-profit/) - [Manufacturing & Logistics](https://fivenines.com/industries/manufacturing-logistics/) - [Small & Mid-Sized Businesses](https://fivenines.com/industries/smb/) - [Insights](https://fivenines.com/resources/) - [Resource Library](https://fivenines.com/resources/) - [Blog](https://blog.fivenines.com) - [About](https://fivenines.com/about/) - [About Five Nines](https://fivenines.com/about/) - [Leadership Team](https://fivenines.com/our-team/) - [Community Impact](https://fivenines.com/community/) - [Partnerships](https://fivenines.com/partnerships/) - [Careers](https://fivenines.com/careers/) - [Company News](https://fivenines.com/news/) Let's Talk Get Support Toggle Menu Toggle Menu Let's Talk Get Support - [Solutions](https://fivenines.com/solutions/) - [Managed IT Services](https://fivenines.com/solutions/managed-it-services/) - [Cybersecurity Services](https://fivenines.com/solutions/cybersecurity/) - [Cloud & Productivity](https://fivenines.com/solutions/cloud-productivity/) - [Projects & Enhancements](https://fivenines.com/solutions/it-projects/) - [Placement Services](https://fivenines.com/solutions/onsite-it/) - Industries - [Finance & Banking](https://fivenines.com/industries/banking/) - [Healthcare](https://fivenines.com/industries/healthcare/) - [Legal](https://fivenines.com/industries/legal/) - [Nonprofit](https://fivenines.com/industries/non-profit/) - [Manufacturing & Logistics](https://fivenines.com/industries/manufacturing-logistics/) - [Small & Mid-Sized Businesses](https://fivenines.com/industries/smb/) - [Insights](https://fivenines.com/resources/) - [Resource Library](https://fivenines.com/resources/) - [Blog](https://blog.fivenines.com) - [About](https://fivenines.com/about/) - [About Five Nines](https://fivenines.com/about/) - [Leadership Team](https://fivenines.com/our-team/) - [Community Impact](https://fivenines.com/community/) - [Partnerships](https://fivenines.com/partnerships/) - [Careers](https://fivenines.com/careers/) - [Company News](https://fivenines.com/news/) # What Does Anthropic's AI Security Expansion Mean for Hospitals? [Five Nines Team](https://blog.fivenines.com/author/five-nines) : Oct 13, 2026, 6:00:01 AM 5 min read [Cybersecurity](https://blog.fivenines.com/topic/cybersecurity) [Healthcare](https://blog.fivenines.com/topic/healthcare) [Tech Trends](https://blog.fivenines.com/topic/tech-trends) [Compliance](https://blog.fivenines.com/topic/compliance) ![What Does Anthropic's AI Security Expansion Mean for Hospitals?](https://blog.fivenines.com/hubfs/Blog%20banner%20Discovery%20Got%20Cheap@1x.png) *The 60 Second Version* - **Finding flaws is no longer the bottleneck.** AI can surface thousands of confirmed vulnerabilities in months, so the constraint moves to who fixes them and how fast. - **The safeguards are real, and so is the capability.** Anthropic's own testing shows the same model blocked on every offensive task for the public can complete about two-thirds of them with restrictions lifted. - **Most hospitals should not apply. They should prepare.** The program asks for proof of security controls and a team to use it, which a hospital with one or two IT staff rarely has. - **Your patch queue is now your perimeter.** Asset visibility, risk-ranked remediation, and compensating controls for devices you cannot patch decide how exposed you are. Anthropic just opened its most capable vulnerability-hunting AI to far more security teams, and its early partners say it found flaws months or even years faster than they could alone. For hospitals, that means the list of known weaknesses in your software is about to grow faster than your team can fix it. It's 6:40 on a Monday morning. The overnight vulnerability scan is waiting in your inbox, and it runs longer than your last board packet. Page one flags the imaging workstation in radiology, still running an operating system its vendor stopped supporting. Last quarter's report flagged it too. So did the one before that. Down the hall, your only IT person is resetting a night-shift nurse's password. ***Discovery just got cheap. Repair didn't.*** Here's how cheap. In four months, partners in Anthropic's Project Glasswing found at least 129,000 verified software vulnerabilities, according to [Anthropic's announcement](https://www.anthropic.com/news/cyber-verification-program). More than 33,000 have been rated critical or high severity so far, counting Anthropic's own open-source scanning, and Anthropic estimates the true impact is at least five times higher. Every one of those flaws lives in software someone still has to patch. Hospitals run a lot of that software, on a lot of machines that are hard to take offline. And yes, regional hospitals can apply for access to this AI themselves. Most shouldn't. Here's why, and what to do instead.   ## What Did Anthropic Actually Announce? > **Short answer:** On October 6, 2026, Anthropic merged Project Glasswing and its earlier verification program into one Cyber Verification Program with three tiers. Verified security teams get Claude Opus 5.5, Sonnet 5.5, and Mythos 5.1 with fewer cyber blocks, scaled to the work they are approved to do. Anthropic's publicly available models carry conservative cyber safeguards that block most security work, because the same skills that defend a network can attack one. The program loosens those blocks for vetted defenders. It does not remove them for everyone. | Tier | What it allows | Who can apply | Review time | | --- | --- | --- | --- | | Defense Access | Security operations, incident response, malware reverse engineering, vulnerability validation | Security teams at companies, nonprofits, universities, and government; critical infrastructure operators of any size, including regional hospitals; smaller security firms; open-source maintainers; qualifying individual researchers | A few days | | Red Team Access | Adds authorized penetration testing against systems the organization has permission to test | Organizations only | A few weeks | | Specialized Access | Fewest blocks, for testing safety-critical systems such as power grids, aviation software, and interbank infrastructure | A small number of organizations vetted with the U.S. government | In-depth review with the U.S. government |   ## Why Does AI Vulnerability Discovery Matter to a Hospital? > **Short answer:** Hospitals run older systems, vendor-locked medical devices, and lean IT teams. When AI makes finding flaws faster for defenders and researchers, the list of known, published weaknesses in that software grows faster too, and every unpatched one becomes easier to target. Most hospital software is not written in house. It comes from EHR vendors, imaging vendors, device makers, and open-source libraries buried inside all of them. When researchers using AI find a flaw in that software, the fix depends on the vendor shipping a patch and the hospital installing it. That second step is where hospitals stall. A patch on a clinical workstation needs a maintenance window. A patch on a medical device can need vendor approval. Some devices cannot be patched at all. Even the organizations inside Anthropic's program show the gap. Anthropic reports that fewer than half of its Glasswing partners disclosed how many of their findings they had patched, often because the fixes were still in progress. If well-resourced security teams are still working through the backlog, a hospital with two IT generalists will be too. > ### The Hidden Risk > > A vulnerability report you cannot act on is not protection. It is a record that you knew.   ## Is the Capability Real if the Public Version Blocks It? > **Short answer:** Yes. In Anthropic's own CyScenarioBench testing, Claude Opus 5.5 with Red Team Access hit no blocks and completed 34 of 50 multi-stage cyber tasks, matching its 67.6% success rate with no safeguards. Without program access, every attempt was blocked at the first prompt. That result cuts both ways. Anthropic's safeguards work: Defense Access still blocked 46 of 50 of the same offensive scenarios. Even Red Team Access still blocks actions like deploying ransomware or damaging physical systems. Vetting controls who gets the less restricted versions of Anthropic's models. Vetting does not set the pace for every AI model on the market. The responsible reading for a hospital leader is simple. Assume the time between a flaw being published and someone trying it on your network is shrinking, and plan your patching around that.   ## Should a Hospital Apply to the Cyber Verification Program? > **Short answer:** Anthropic names regional hospitals as an example of who qualifies for Defense Access, but most critical access hospitals and clinics are not positioned to use it. Anthropic asks applicants for proof of required security controls and requires data retention for monitoring. Without a security team to run the work, access adds little. The program is built for people who already do security work every day: SOC analysts, incident responders, penetration testers. A hospital with one or two IT generalists would be applying for a tool nobody has hours to operate. The better question is who on your side already has that capacity. That might be an internal security team at a larger system, or a managed security partner with a staffed security operations center. > ### What Leadership Misses > > The question for 2027 is not whether your hospital can find its weaknesses. It is how many days each one stays open after you do.   ## What Is the Five Nines Find-to-Fix Model? > **Short answer:** The Five Nines Find-to-Fix Model is a four-stage way for hospital leaders to judge where their vulnerability program actually stands, from Blind to Closed Loop. It measures remediation, not scanning, because AI has made scanning the easy part. **The Five Nines Find-to-Fix Model** | Stage | What it looks like on the floor | What usually triggers the next step | | --- | --- | --- | | 1. Blind | Nobody can produce a current list of every device on the clinical network | A cyber insurance questionnaire or an incident at a peer hospital | | 2. Buried | Scans run, but the same critical findings show up report after report | An auditor or insurer asks for proof of remediation | | 3. Ranked | Findings are prioritized by exploitability and clinical impact, each with an owner and a date | Leadership asks for a weekly view of exposure | | 4. Closed Loop | Fixes are verified, evidence is kept for HIPAA, and devices that cannot be patched are segmented and monitored | Time-to-fix becomes a number the board sees |   ![Graphic 1 Find-to-Fix Model@1x](https://blog.fivenines.com/hs-fs/hubfs/Graphic%201%20Find-to-Fix%20Model@1x.png?width=1200&height=1500&name=Graphic%201%20Find-to-Fix%20Model@1x.png)   > ### Five Nines Insight > > When finding a flaw costs almost nothing, the value of a security partner moves entirely to what happens after the finding.   ## What Does Five Nines Usually Find Before a Hospital Patch Program Stalls? > **Short answer:** Five Nines Technology Group engineers see the same handful of patterns in hospital environments: devices nobody inventoried, findings with no owner, flat networks, and clinical schedules that leave no patch window. None of them are scanning problems. All of them slow the fix.   - **Vendor-installed devices outside the inventory.** Equipment set up by an outside vendor often never lands in the hospital's asset list, so scans miss it or nobody claims it. - **Repeat findings with no owner.** The same critical item appears across several reports because no one person is accountable for closing it. - **Clinical and business systems on one network.** When medical devices share a network with email and guest Wi-Fi, one weak device exposes everything around it. - **Unsupported systems held in place by vendor certification.** Imaging and lab workstations stay on old operating systems because upgrading would void the vendor's support. - **No agreed patch window.** In a 24/7 care setting, patching waits until someone negotiates downtime with nursing leadership, and that conversation keeps getting pushed.   ## What Should a Hospital Ask Its IT or Security Provider Now? > **Short answer:** Ask how they rank and close vulnerabilities, not how many they find. Five questions reveal whether a provider can keep pace with AI-speed discovery or will hand you longer reports. 1. How do you decide which vulnerabilities get fixed first, and does clinical impact factor in? 2. What is your typical time from a critical finding to a verified fix? 3. What do you do with devices that cannot be patched? 4. Who watches our environment at 2 a.m., your analysts or an automated alert queue? 5. If you use AI tools in security work, how is our data handled and retained?   ## How Long Do Your Known Vulnerabilities Stay Open? If your last scan report looked a lot like the one before it, you already have your answer. A Five Nines Find-to-Fix assessment identifies: - Every device on your clinical and business networks, including vendor-installed equipment - Which open findings carry real exploit risk and clinical impact - Repeat findings and why they keep coming back - Systems that cannot be patched and the controls they need instead - Where your hospital sits on the Find-to-Fix Model today You leave with a ranked remediation plan, an owner for each item, and the evidence trail HIPAA and your cyber insurer will ask for. The scanner will keep getting faster. What matters is how fast the fix follows.   ![Graphic 2 Discovery vs Repair Gap@1x](https://blog.fivenines.com/hs-fs/hubfs/Graphic%202%20Discovery%20vs%20Repair%20Gap@1x.png?width=1200&height=1300&name=Graphic%202%20Discovery%20vs%20Repair%20Gap@1x.png)   ## Frequently Asked Questions What is Anthropic's Cyber Verification Program? It is a trusted access program that gives verified security professionals Claude models with fewer cybersecurity restrictions than the public versions. Since October 6, 2026, it has three tiers: Defense, Red Team, and Specialized Access, each with its own verification and security requirements. Can a hospital apply to the Cyber Verification Program? Yes. Anthropic names regional hospitals among the organizations eligible for Defense Access, which covers security operations, incident response, and malware analysis. Applicants must show required security controls, and data retention is required so Anthropic can monitor for misuse. Anthropic's public models can still help with reviewing code, patching known issues, and triaging security alerts. Most smaller hospitals will get more value from a partner that already runs this kind of work. Does AI make hospitals easier to attack? AI makes finding software vulnerabilities faster, for defenders and for anyone else with capable tools. That can shorten the time between a flaw becoming known and someone trying it. The practical defense is faster, risk-ranked patching and monitoring, not avoiding AI. What should a hospital do about devices that cannot be patched? Isolate them on their own network segment, limit what can talk to them, and monitor them continuously. Document those compensating controls, because auditors and cyber insurers will ask how unpatchable systems are protected. How does Five Nines help hospitals keep up with AI-speed vulnerability discovery? Five Nines Technology Group combines vulnerability management, patch management, and 24/7 monitoring from an in-house security operations center. The focus is closing findings by risk and clinical impact, with evidence kept for HIPAA, so known weaknesses do not sit open. ## Related Blog Posts ## Let's get in touch   [**(402) 817-2630**](tel:402-817-2630) [help@fivenines.com](mailto:help@fivenines.com) [Lincoln, NE](https://fivenines.com/contact/lincoln) [Omaha, NE](https://fivenines.com/contact/omaha) [Kearney, NE](https://fivenines.com/contact/kearney) [Central City, NE](https://fivenines.com/contact/central-city) [St. Louis, MO](https://fivenines.com/st-louis) ![cyberverify-1](https://blog.fivenines.com/hs-fs/hubfs/cyberverify-1.png?width=110&height=110&name=cyberverify-1.png)![aicpa](https://blog.fivenines.com/hs-fs/hubfs/aicpa.png?width=110&height=110&name=aicpa.png) ### Are we a good fit? [![Schedule Consultation](https://no-cache.hubspot.com/cta/default/1857420/interactive-196258273705.png)](https://blog.fivenines.com/hs/cta/wi/redirect?encryptedPayload=AVxigLL0Gih9erDjl%2FOA%2Fx9QQl%2FeMUTmSlWPLwVwu27DclIRoRDzjydSqrVh5kcWyFHbkdPzIZRYa4jBYYkyaToisHTCdx%2FOcmL03%2BL%2FNxhG8uB9iny4fwXvUExaK7mE28EHAbWTg8Msh6Hk9NhiMTlAE2s0L8Vg82akNTthVeF0j6WTKLtBlhzdD6oQ03o4gc3AcrsjrA%3D%3D&webInteractiveContentId=196258273705&portalId=1857420)   - [Privacy Policy](https://fivenines.com/privacy-policy/) - [HTML Sitemap](https://fivenines.com/html-sitemap/) © 2026 Five Nines Technology Group | 5617 Thompson Creek Blvd Lincoln, NE 68516 [Facebook](https://www.facebook.com/gonines)[Linkedin](https://www.linkedin.com/company/five-nines-technology-group/) [Blog](https://blog.gonines.com/?__hstc=143714730.45718742b0726c421d8592d7ea71f58b.1757514685996.1758029308186.1758134756251.4&__hssc=143714730.31.1758134756251&__hsfp=1745665186)   [Client Support](https://fivenines.com/client-login/) Return to top ```json { "@context" : "https://schema.org", "@type" : "FAQPage", "mainEntity" : [ { "@type" : "Question", "acceptedAnswer" : { "@type" : "Answer", "text" : "It is a trusted access program that gives verified security professionals Claude models with fewer cybersecurity restrictions than the public versions. Since October 6, 2026, it has three tiers: Defense, Red Team, and Specialized Access, each with its own verification and security requirements." }, "name" : "What is Anthropic's Cyber Verification Program?" }, { "@type" : "Question", "acceptedAnswer" : { "@type" : "Answer", "text" : "Yes. Anthropic names regional hospitals among the organizations eligible for Defense Access, which covers security operations, incident response, and malware analysis. Applicants must show required security controls, and data retention is required so Anthropic can monitor for misuse. Anthropic's public models can still help with reviewing code, patching known issues, and triaging security alerts. Most smaller hospitals will get more value from a partner that already runs this kind of work." }, "name" : "Can a hospital apply to the Cyber Verification Program?" }, { "@type" : "Question", "acceptedAnswer" : { "@type" : "Answer", "text" : "AI makes finding software vulnerabilities faster, for defenders and for anyone else with capable tools. That can shorten the time between a flaw becoming known and someone trying it. The practical defense is faster, risk-ranked patching and monitoring, not avoiding AI." }, "name" : "Does AI make hospitals easier to attack?" }, { "@type" : "Question", "acceptedAnswer" : { "@type" : "Answer", "text" : "Isolate them on their own network segment, limit what can talk to them, and monitor them continuously. Document those compensating controls, because auditors and cyber insurers will ask how unpatchable systems are protected." }, "name" : "What should a hospital do about devices that cannot be patched?" }, { "@type" : "Question", "acceptedAnswer" : { "@type" : "Answer", "text" : "Five Nines Technology Group combines vulnerability management, patch management, and 24/7 monitoring from an in-house security operations center. The focus is closing findings by risk and clinical impact, with evidence kept for HIPAA, so known weaknesses do not sit open." }, "name" : "How does Five Nines help hospitals keep up with AI-speed vulnerability discovery?" } ] } ```