From the frontier artificial intelligence front

• The Wall Street Journal reports,
  • “Rogue AI models have hacked companies and tried to break into government websites. Now one has submitted a fake tip for an unsolved crime.
  • “Police in Philadelphia say Anthropic alerted them this week that one of its artificial-intelligence models submitted a false homicide tip through PhillyUnsolvedMurders.com. 
  • In a blog post, Anthropic said the fake tip was among a group of rogue actions the company uncovered recently. The rogue activity also involved websites run by U.S. government agencies at the federal, state and local levels. The company said it has briefed the White House on these cases and notified each agency involved, and plans to report new instances of unintended AI behavior as it uncovers them.
  • “While these cases had minimal impact, we do not want to diminish the findings, because the same behaviors could do far more harm as models become more powerful,” Anthropic said in the blog post.
  • “The false-report incident happened in July. Anthropic didn’t realize it until September.”
• and
  • “Anthropic edged to the front of the AI race earlier this year with cutting-edge models and a coding tool that corporate America loved. By some measures, OpenAI is now hot on its heels.
  • Demand for Anthropic’s Claude Code was so strong in the spring that the company faced frequent outages and a computing crunch. 
  • “But businesses grappling with mounting costs as more employees experimented with artificial intelligence realized they didn’t always need the most advanced options for most tasks. Companies, which now use a variety of models, increasingly turned to low-cost, open-weight models from China for some tasks. And some Claude users balked at guardrails Anthropic put on its powerful Fable 5 model when it was released in early June.
  • “OpenAI stoked the price war this summer with the release of its GPT-5.6 lineup—Sol, Terra and Luna—giving customers access to models with varying capabilities, including some that are less expensive to run. Its Codex coding product and other business tools are now gaining ground, and it has released additional iterations of its cost-efficient models.
  • “An analysis from OpenRouter, a startup that allows developers to access different models, found that among some 120,000 companies that use Anthropic and OpenAI tools, the share of spending was roughly even between the two AI giants in September. At the beginning of the year, Anthropic commanded three-quarters of that total. OpenRouter said its data largely represents spending by AI-native startups as well as some slightly older tech companies and large enterprises.”
• and
  • “Mistral said it would soon release a new open-weight AI model that it claimed would rank among the best globally, part of efforts to grow its appeal, lock in more customers and better compete with U.S. rivals.
  • “The French artificial-intelligence startup said it was launching a public preview of Mistral Large 4, dubbed Le Chonk, on Tuesday so it could work with developers, cybersecurity experts and state authorities to evaluate how the model behaves. It plans to release the weights, or the trained parameters of the model, on Oct. 27.
  • “Since its early days, Mistral has been a staunch advocate of open-weight models that are publicly available for anyone to download and modify, saying they give users the flexibility to tailor them to their needs while protecting their data and intellectual property.”

From the cybersecurity policy and law enforcement front

• The Wall Street Journal reports,
  • “A new White House task force will have 120 days to report on the risks and opportunities of artificial intelligence—and determine the federal government’s responsibility—according to Jay Clayton, the director of national intelligence who effectively becomes the administration’s AI czar.
  • “The president asked that a group be put together that was going to ensure exactly what he said,” Clayton said, that “we stay the leaders in superintelligence, and that the interests of the American people are put first.” 
  • “Clayton will chair the “Super Intelligence Force” which takes President Trump’s preferred name for AI: super intelligence or SI.
  • “The risk of not being first is high,” Clayton said. “Not being first increases the identified, and unidentified, risks, particularly from our adversaries. Being first will better enable us to address those risks on behalf of the American people,” he added.” * * *
  • “The task force will have three vice chairs: Emil Michael, undersecretary of war for Research and Engineering, who will be responsible for coordination with AI companies; Scott Kupor, director of the Office of Personnel Management, who will be responsible for coordination with critical infrastructure providers and other industry stakeholders; and Andrew Ferguson, chairman of the Federal Trade Commission, who will oversee consumer interests and groups advocating for people of faith, children, privacy, freedom of speech, freedom of expression, and freedom of association.”

• Cyberscoop relates,
  • “Federal government contractors that handle sensitive information could soon face a “sea change” in rules about how they protect that information and report when it has been part of a breach.
  • “Pending federal regulations on the handling of “controlled unclassified information,” or CUI, a category of sensitive data that falls short of classified — including people’s personal information such as Social Security numbers, information that could expose vulnerabilities in critical infrastructure and more — could arrive as soon as the end of this year, but likely no later than the end of President Donald Trump’s term, attorneys who specialize in federal procurement rules said. The changes are part of a larger overhaul of federal contracting rules.
  • As currently written, the proposed CUI rules mandate that unauthorized access of such data, including as a result of a cyberattack, would have to be reported to the federal government within 72 hours of discovery. The rules are a companion for most federal agencies to existing  Defense Department rules that cover the same subject. The 72-hour notice also is deliberately aligned with forthcoming rules from the Cybersecurity and Infrastructure Security Agency (CISA) for critical infrastructure owners and operators to report major cyber incidents under the Cybersecurity Incident Reporting for Critical Infrastructure Act (CIRCIA).” * * *
  • The other major aspect of the proposed rule is that contractors handling CUI would have to abide by certain cybersecurity standards set by the National Institute of Standards and Technology known as SP 800-171, which [Trayce] Howard and her Wiley [Rein] coauthors said would be imposed “on a broader group of contractors than ever before.” Some contractors might also have to meet other cybersecurity requirements, too.
  • “Another thing that will affect contractors is they’re going to have to flow down these requirements to their subcontractors,” she said. “So they’re going to have to identify what CUI am I giving to my subcontractors, and do some sort of oversight to make sure that their subcontractors are appropriately handling CUI.”

• Cybersecurity Dive tells us,
  • “U.S. government agencies are failing to prepare for the transition to post-quantum encryption, creating the risk that their sensitive data will still be vulnerable to decryption by quantum computers whenever those machines are deployed, according to the Government Accountability Office.
  • Experts are divided over when the advent of cryptographically relevant quantum computers (CRQC) will occur, but GAO bluntly declared in a report published Tuesday [October 6]  that “threats from a CRQC could have devastating impacts to federal systems reliant on vulnerable cryptography.” * * *
  • “In a version of the report not publicly released, GAO recommended 89 actions that CISA and other agencies could take to increase their preparedness. Of the 23 agencies to which GAO assigned recommendations, only 12 fully agreed to implement them. “We continue to believe that all of the recommendations made in the sensitive report are warranted,” GAO said.”

• The Wall Street Journal notes,
  • A bipartisan group of federal lawmakers is pressing Congress to increase mental-health funding for personnel at U.S. Cyber Command in the wake of several suicides this summer at the command and the National Security Agency. In separate letters, members of the House Armed Services cyber subcommittee and Maryland lawmakers urged leaders of the House and Senate Appropriations panels to back an $11 million request for “High Performance Team Training” at the command. (The Baltimore Banner)
• and
  • German officials said they infiltrated ransomware group Qilin and detained a 28-year-old Russian national believed to be one of the group’s leaders. The arrest follows months of covert monitoring of Qilin’s activities, the official said. The group is alleged to have demanded more than $3 billion in ransom payments over the past two years from targets around the world. (LIGA)

• The New York Times adds,
  • “The F.B.I. has arrested a Canadian cybersecurity expert in Pennsylvania on suspicion of carrying out the computer hack that stole sensitive information about thousands of F.B.I. employees, according to two people familiar with the matter.
  • “The theft of reams of personal data last month by a criminal hacking group that calls itself ShinyHunters left the nation’s premier law enforcement agency racing to respond to the investigators and the victims of one of the worst breaches of government data in recent years.
  • “The arrested man is Edward Dubrovsky, who has helped firms victimized by hackers and is now considered by federal authorities to be a primary co-conspirator in the ShinyHunters attack on F.B.I. data, the people said.
  • “Mr. Dubrovsky is the author of a book published last month titled “Cyber Extortion Strategic Response.” While the details of the charges against Mr. Dubrovsky remain under seal in Texas, federal court records in Philadelphia show that he was charged with conspiracy to commit extortion. Mr. Dubrovsky is expected to be transported to Texas to face the charges.”

• Per a Health and Human Services Department news release,
  • “The U.S. Department of Health and Human Services (HHS), Office for Civil Rights (OCR) today announced a settlement with Linda L. Shen d/b/a Shen Smiles, P.C. (Shen Smiles), a Pennsylvania-based healthcare provider that offers dental services, for potential violations of the Health Insurance Portability and Accountability Act of 1996 (HIPAA) Privacy Rule. The settlement resolves an investigation of a complaint alleging a failure to provide timely access to an individual’s protected health information (PHI).”
  • “The Notice of Proposed Determination may be found here.
  • “The Resolution Agreement may be found here [PDF, 161 KB].”

From the cybersecurity breaches and vulnerabilities front

• HIPAA Journal reports,
  • “The scale of a 2025 data breach involving electronic protected health information stored on Oracle Health’s legacy Cerner servers is becoming clearer. While not independently verified, Bloomberg reports that the Texas Attorney General released information suggesting that the protected health information of almost 20 million individuals was compromised in the incident. Oracle Health has not publicly disclosed how many Cerner clients were affected, or the total number of individuals affected.”

• Security Week relates,
  • “Healthcare organizations Clover Health Investments and AngMar Management Services are notifying more than 250,000 people that their information was stolen in separate data breaches.
  • “Jersey City, New Jersey-based Clover Health Investments was hacked in early July, after attackers used social engineering to compromise three non-managerial health plan employee accounts.
  • “The incident resulted in the theft of personally identifiable information (PII) and protected health information (PHI), the company said in an SEC filing in July.
  • “The potentially affected information included names, dates of birth, insurance identifiers, and account identification numbers, the company revealed.
  • “In mid-September, Clover Health Investments told the US Department of Health and Human Services (HHS) that 138,677 people were affected. HHS added the company to its data breaches portal last week.
  • “Mansfield, Texas-based AngMar Management Services identified suspicious activity on its systems in mid-July and confirmed in early September that hackers stole patient PII and PHI.
  • “AngMar Management Services provides business operations, administration, and support network management for home health and hospice care providers.”

• The American Hospital Association News tells us,
  • “The FBI and Secret Service released an advisory Oct. 6 warning of ongoing activities by FortiBleed, a global credential-compromise campaign used by threat actors to target internet-facing Fortinet FortiGate firewalls and virtual private network gateways at critical infrastructure organizations. The agencies said the campaign exploits reused or leaked credentials and legacy password storages, allowing threat actors to harvest and decode authentication data at scale. Impacted organizations could be locked out of their systems if threat actors disable accounts or change passwords and would require remediation actions beyond typical patching and password resets. The advisory includes indicators of compromise and mitigation actions to defend against potential attacks.”
• and
  • “A joint advisory released Oct. 8 by U.S. and international agencies warns of Chinese government-linked cyber threat actors using automated and hands-on hacking tools to steal sensitive data from critical infrastructure organizations, including healthcare, across the U.S. and abroad. The advisory said that the threat actors are enabled by the Integrity Technology Group, a China-based company with links to the Chinese government. The agencies said the threat actors are using automated scanning tools, large-scale botnets and hands-on exploitation techniques to attack Microsoft Exchange servers and virtual private network software. The advisory includes recommendations to help mitigate risk from potential attacks.” 

• Cyberscoop adds.
  • “The Justice Department and FBI announced that they had seized two hacking tools connected to the Chinese government-linked group Flax Typhoon and a China-based company that the U.S. government has repeatedly taken action against, including with a new multi-agency advisory Thursday.
  • “The domain name seizures were meant to deny hackers access to the vulnerability scanning tool Microscan and the spearphishing tool FishHub created by the Chinese firm Integrity Technology Group, which the United States sanctioned last year. The U.S. government in 2024 also made the company the focus of a takedown operation, saying it was behind a massive botnet.
  • “U.S. agencies paired the announcement of the seizures with an Oct. 8 advisory from the FBI, Cybersecurity and Infrastructure Security Agency and National Security Agency.”

• CISA added one known exploited vulnerability to its catalog this week.
  • October 4, 2026
    • CVE-2026-88779 Citrix NetScaler Improper Restriction of Operations within the Bounds of a Memory Buffer Vulnerability
      • Cyber Press discusses this KVE here.

• Bleeping Computer notes,
  • “Attackers are exploiting a maximum-severity vulnerability in SonicWall SMA1000 appliances (CVE-2026-102255) that was patched on Tuesday, three days ago.
  • “Tracked as CVE-2026-102255, the flaw affects the Appliance WorkPlace interface on SMA1000 6210, 7210, and 8200v models, but does not affect the SMA 100 Series product line or SSL-VPN running on SonicWall firewalls.
  • “By abusing this path, a remote unauthenticated attacker could potentially exploit this vulnerability to direct the appliance to issue requests on their behalf and reach internal functionality and perform unauthorized operations,” SonicWall explained.”
• and
  • “Citrix has warned IT administrators to patch systems immediately against a new critical vulnerability affecting NetScaler ADC networking appliances and NetScaler Gateway secure remote access solutions.
  • “Tracked as CVE-2026-107406, this flaw stems from a memory overflow weakness that attackers can exploit to gain remote code execution (RCE) on targeted devices or trigger a denial-of-service state that can cause crashes.
  • “To be vulnerable, NetScaler ADC and NetScaler Gateway appliances must be configured as a Security Assertion Markup Language (SAML) Identity Provider (IdP) or Service Provider (SP).
  • “We strongly urge affected customers to review the advisory and upgrade impacted NetScaler instances to the recommended versions as soon as possible,” the company said. “As of the publication of the bulletin, Citrix is not aware of any unmitigated exploits of this vulnerability.”

• Per Cybersecurity Dive,
  • “Hackers may be using AI more frequently and in more ways, but their top priority is still subverting identity controls, the identity-security firm Okta said in a report published on Wednesday.
  • “The report included data about the frequency of different kinds of account takeovers and a reality check to enterprises that AI shouldn’t fundamentally change their calculus when it comes to identity management.
  • “In a separate report also published on Wednesday, Okta detailed how hackers are trying to bypass strong authentication controls to sneak into accounts, underscoring the need for vigilance among both system administrators and regular employees.”

• Security Boulevard adds,
  • “An analysis of 150 billion transactions flowing through Domain Name System (DNS) servers published today by EfficientIP finds that malware has overtaken phishing as the threat vector now being detected most.
  • “The latest DNS Threat Intelligence Report from EfficientIP finds malware has climbed from fourth place in the second half of 2025 to first in daily activity, having roughly quadrupled between a January and an April peak of 32.6 million instances.
  • “Yaelle Harel, principal security market strategist for EfficientIP, said that while there is no conclusive evidence, that level of increase suggests cybercriminals may now be leveraging artificial intelligence (AI) tools to both create malware and launch attacks faster using newly acquired IT infrastructure resources.”

From the ransomware front

• Info Security Magazine reports,
  • “Ransomware attacks reached their highest ever quarterly volume in the third quarter of 2026, according to an analysis by Comparitech.
  • “The firm identified a total of 2627 claimed attacks during the three-month period from July to September 2026. This is a 27% increase on the previous quarter, Q2 2026, and a 61% rise compared to Q3 2025.
  • “The finance and technology sectors experienced the biggest growth in ransomware incidents in Q3 compared to Q2 2026, up by 72% and 70%, respectively.
  • “Other critical sectors faced a significant rise in attacks, including education (up 50%), healthcare (39%), government (36%) and utilities (32%).
  • Of the 2627 attacks claimed by ransomware groups in Q3, 247 have been confirmed by the entity involved, Comparitech found.
  • “Rebecca Moody, head of data research at Comparitech, said that the figures are highly unusual, and represent a significant cause for concern.: * * *
  • “A possible explanation for the surge is developments in AI technology, which is enabling ransomware attackers to increase the scale and speed of campaigns, as well as their effectiveness.” * * *
  • “The report highlighted an increase in attackers using triple extortion tactics on victim organizations. In this model, in addition to encrypting systems and exfiltrating data, threat actors also target individuals impacted in the attack.
  • Moody added that this tactic further demonstrates that paying a ransom is no guarantee that the attacker will keep to their word regarding deleting stolen data.

• Recorded Future News relates,
  • “The University of Illinois Chicago (UIC) recently discovered a ransomware attack that limited access to some systems at its College of Medicine.  
  • “A spokesperson for the university told Recorded Future News the hackers were able to steal some information held on the college’s servers and an investigation is underway to determine whether “any personal, research or academic information was compromised.” 
  • “As a result of this ransomware event, some College of Medicine systems were temporarily unavailable,” the spokesperson said. “However, all affected systems have since been restored. The university’s main network was not affected, and there was no impact on patient care delivery at UI Health.” * * *
  • “The attack was claimed last week by a ransomware gang known as Booba, which claimed to have stolen 344 gigabytes of data. The group emerged at the end of July and has already claimed 49 attacks. 
  • “SentinelOne’s Brett Williams said the group appears to be a rebrand of the Frag ransomware based on the leak site’s style and negotiation flow. Encrypted files are renamed with the .booba extension and there appears to be variants for Linux as well as Windows.  
  • “The group has targeted several companies and small county governments, including Merrimack County, New Hampshire.” 

• Cyberscoop tells us,
  • “Authorities accuse the owner of a so-called ransomware remediation company of swindling clients victimized by ransomware attacks into paying the company inflated fees under false pretenses.
  • “Zohar Pinhasi, owner and operator of MonsterCloud, claimed he could decrypt and recover victims’ data with specialized, proprietary tools and avoid paying cybercriminals. Yet, no such tool existed, the Justice Department said Wednesday [October 7]. 
  • “Pinhasi allegedly used MonsterCloud clients’ fees to pay off cybercriminals and recover encrypted data without telling clients ransom payments were made on their behalf. The dual “U.S.-Israeli national is accused of charging hundreds of clients more than $19 million and paying more than $8 million in ransom payments during a five-year period ending in 2023.
  • “It baffles me that companies fall for this,” Jon DiMaggio, principal researcher at Arkem Cyber, told CyberScoop. “I mean, they just got hacked and they go straight to a shady source.”

From the cybersecurity business and defenses front,

• Cybersecurity Dive discusses what the C-suite needs to know about AI governance
  • “As AI adoption surges, executives are learning tough lessons about security, oversight and accountability.”

• The American Hospital Association News tells us,
  • The National Security Agency Oct. 8 released guidance for critical infrastructure organizations that recommends adopting high-impact zero trust strategies to improve cybersecurity for operational technology. For healthcare, such technology includes tools that manage energy control, HVAC, life-safety systems, door access controls, physical security systems, alarms and medical gasses. Zero trust is a security principle that assumes a compromise is possible or already present and continuously verifies users and devices. The guidance also highlights recent nation-state attacks, including the People’s Republic of China’s prepositioning of disruptive malware known as Volt Typhoon on critical infrastructure entities that healthcare depends on, such as energy. The guidance also explains how threat actors’ use of artificial intelligence is making cyberattacks more sophisticated and potentially more severe.

• Security Week points out,
  • “All enterprises conduct security awareness training for their employees. But whether this has tangible benefits is debatable.
  • “Empirical evidence suggests that security awareness training isn’t working – successful attacks keep increasing. But opinions on the efficacy of awareness training range from it doesn’t work to it does work, with sometimes, perhaps and depends between the two extremes.
  • “Awareness training focuses on two tasks: to reduce the effect of bad judgment turning an employee into an insider threat; and to harden employees against falling to malicious social engineering. We focus on social engineering.” * * *
  • “Sanny Liao, co-founder and CPO at Fable Security, says bluntly, “For the most part, security awareness training as done today does not work. Social engineering continues to succeed because attackers exploit context, timing, and psychology, while most training remains generic, infrequent, and removed from the moments when employees are actually making decisions.”
  • “She hints at a novel approach. “One place the industry can look for inspiration is adtech. Marketers have gotten remarkably good at changing behavior by delivering the right message at the right time based on context. Security awareness has largely done the opposite by giving everyone the same training at the same time, regardless of the decisions they’re making or the risks they face. When organizations change behavior instead of simply raising awareness, employees stop being viewed as the weakest link and start becoming an active part of the organization’s security defenses.”

• Per a NIST news release,
  • “For small businesses who are often confronted with limited resources, knowing how to get started and where to find support with planning, implementing, or evaluating a cybersecurity risk management strategy can be challenging — sometimes making cybersecurity feel like an insurmountable hurdle. However, there is good news.
  • “Many non-profit organizations across the United States have created programs for their local small business communities (in addition to local schools, municipal governments, and non-profits) that often extend beyond general cybersecurity education to include services like conducting risk assessments and penetration tests, performing compliance consulting services, helping with incident preparedness and response measures, and more.
  • “The NIST Small Business Cybersecurity Support Program Finder brings these programs together into one tool, connecting small businesses with local non-profit community partners who help with cybersecurity risk management.”

• Here is a link to Dark Reading’s CISO Corner.