The Week in Review
- The Wall Street Journal reports,
- “The inevitability—the invincibility—of Big Tech felt diminished Thursday with Jensen Huang’s ouster from the company he founded and built into the backbone of the AI boom.
- “As his power slipped away at Nvidia NVDA , the court of tech royalty was being dealt an uncomfortable truth about their new place in the world: Big Tech isn’t so big anymore nor, perhaps, so unbeholden. * * *
- In the span of a little more than a week, the titans lost their grip on political power in Washington, D.C., followed by one of their biggest stars.” * * *
- “The power of tech billionaires was also on trial during the midterm elections, where AI data centers were a major campaign issue. For the past two years, the likes of Musk, Huang and the broader AI leadership have flourished under President Trump, who looked to them for advice and attention.
- After riding a blue wave into power, Democrats are signaling that tech will face greater scrutiny come January.
- The New York Times discusses,
- “Microsoft’s Global Outage and Fixes: What We Know
- “Disruptions to the company’s email, messaging and other products lasted about 36 hours, hampering companies and services around the world.”
- Was this a cyberattack?
- “Microsoft said in a statement on Friday that “for now, we can confirm that there is no indication this is the result of malicious or cyber activity.” It added that the problem was “a technical issue not related to cyber or A.I.” agents.
- Was the F.A.A. ground stop linked to Microsoft?
- “The company said it had no indication that a Federal Aviation Administration ground stop issued late Friday afternoon, which has affected air travel nationwide, was related to the outage.
- So why were Microsoft’s systems not working?
- “The company has yet to provide a clear answer, although it has promised to make public a full report identifying a “root cause.”
- “Microsoft’s Global Outage and Fixes: What We Know
- NPR lets us know,
- “The nationwide ground stop that paralyzed the U.S. aviation system on Friday has ended, according to the Federal Aviation Administration.
- “The rare stop, which lasted about four hours, grounded thousands of flights at all U.S. airports in one of the biggest disruptions of the nation’s air travel system since the September 11th attacks.
- “In a Friday night news conference, Transportation Secretary Sean Duffy said that air traffic control facilities had experienced “intermittent” outages of their radar and communications systems beginning earlier in the afternoon. Duffy blamed outdated technology for the failures.
- “We have old switches, old routers. We use 1980s little copper lines. It’s an old system that went down and went down everywhere. It got overloaded,” he said.
- “But when pressed by reporters on exactly why those systems would become overwhelmed on an otherwise normal travel day, Duffy did not have a ready answer.
- “Homeland Security Secretary Markwayne Mullin joined Duffy in the White House briefing room and said there was no evidence the issue was due to an attack.
- “We have seen zero evidence this is a cyber attack,” he said.”
From the cybersecurity policy and law enforcement front,
- Federal News Network reports,
- “Cyber leaders in the next Congress say they’re focused on restaffing the Cybersecurity and Infrastructure Security Agency, extending key cyber information sharing authorities, strengthening the security of critical infrastructure and addressing artificial intelligence security risks.
- “With Democrats retaking both the House and the Senate, the expected leaders of committees with cybersecurity oversight are staking out their priorities for January.
- “Rep. Bennie Thompson (D-Miss), who is expected to retake the reins of the House Homeland Security Committee, said the change in congressional leadership comes at a “pivotal time for CISA” as the agency hires several hundred staff after losing more than 1,000 – roughly one-third of its workforce – since January 2025.” * * *
- “Sen. Maggie Hassan (D-N.H.), who is expected to serve as chairwoman of the Senate Homeland Security and Governmental Affairs Committee, also highlighted CISA staffing as key to her homeland security priorities.” * * *
- “Beyond CISA staffing, lawmakers are also looking to revive or advance partnerships and authorities that are key to how the federal government works with state and local governments, and the private sector, on cyber issues.”
- Cyberscoop adds,
- “The two expected incoming leaders of the Senate and House homeland security panels expressed reservations Tuesday about the Trump administration’s pick to lead the Cybersecurity and Infrastructure Security Agency because he currently works for the oft-controversial tech company Palantir.
- “The Trump administration on Monday announced its intent to nominate Michael LoSapio, chief information security officer for Palantir, to fill the long-vacant Senate-confirmed position of CISA director.
- “The Senate Homeland Security and Governmental Affairs Committee has authority over approving CISA director nominees. A spokesperson for its current chairman, Rand Paul, R-Ky., didn’t respond to a request for comment, but he won’t hold the gavel much longer with Sen. Maggie Hassan, D-N.H., in line to take over as chair in 2027.
- “Hassan and her counterpart on the House side, prospective Homeland Security Committee Chairman Bennie Thompson, D-Miss., both voiced concerns about Palantir’s work at the Department of Homeland Security and with its subagency, Immigration and Customs Enforcement.”
- Cybersecurity Dive relates,
- “The Trump administration’s AI task force is encouraging regulatory agencies not to take enforcement actions against companies that experience cyberattacks resulting from AI overwhelming their defenses.
- “It is clear that the cybersecurity risk environment has changed and organizations that only a short time ago had reasonable protections in place are now more vulnerable,” the new task force said in a statement about its second meeting on Thursday [November 12, 2026]. “As a result, prior standards for taking regulatory action against organizations in response to cyber incidents may be outdated.”
- and
- “The Cybersecurity and Infrastructure Security Agency will focus on understanding risky interdependencies between critical infrastructure, disrupting significant cyber threats and reducing systemic vulnerabilities to promote national resilience, according to a recent internal memo.
- “Nick Andersen, CISA’s acting director, spelled out the cybersecurity agency’s Fiscal 2027 goals and the information requirements for achieving them in a Nov. 2 memo to agency leaders obtained by Cybersecurity Dive. The memo lists the agency’s three top-line priorities — leading national cyber defense, prioritizing the security of key infrastructure and overseeing nationwide resilience efforts — along with associated activities, intelligence-collection needs and mission-support requirements.
- “The intent is to create an organization in which personnel at every level understand where we are focused, what we need to know, why that information matters, and when it must be elevated,” Andersen wrote in the memo, which Nextgov first reported.”
- Per a CISA news release,
- Today [November 9, 2026], the Cybersecurity and Infrastructure Security Agency (CISA) unveiled the updated cyber incident reporting form, an enhanced web-enabled questionnaire available on CISA’s Incident Reporting Portal for reporting cyber incidents. By improving data quality, the more robust form strengthens the effectiveness and efficiency of national cyber defense. For instance, the new form will help CISA correlate related cyber threat activity to identify a larger trend from what otherwise would appear as independent incidents.
- The new cyber incident reporting form is guided by community feedback CISA has received over the years and the agency’s efforts to improve data quality that enables rapid yet thorough threat analysis. For example, questions about the type of incident are more comprehensive and seek more granular information, including whether the incident involved malware, social engineering, hacking techniques or an insider threat. Also, questions about observed activity now incorporate the MITRE ATT&CK® Framework, language about adversary behavior that is familiar to the cyber community.
- “In the team sport of cybersecurity, one organization’s detection is another’s protection,” said Acting Executive Assistant Director for Cybersecurity Chris Butera. “As the national coordinator for critical infrastructure cybersecurity and resilience, CISA is able to leverage such incident reporting to identify emerging cyber campaigns, warn potential targets, and strengthen the Nation’s collective defense.”
- While some entities are required to report to CISA (e.g., federal civilian agencies and certain entities regulated by Transportation Security Administration or the U.S. Coast Guard), we strongly encourage any organization that experiences a cyber incident to report voluntarily.
- For more information, please visit the Frequently Asked Questions webpage.
- Federal News Network informs us,
- “Government cloud service providers are scrambling to adopt new cyber vulnerability management practices mandated in response to how artificial intelligence is helping hackers accelerate attacks on unpatched software bugs.
- “The General Services Administration’s centralized cloud security program, FedRAMP, is requiring certified cloud services to take steps to comply with new Vulnerability Detection and Response (VDR) and Vulnerability Evaluation and Reporting (VER) rules by early December.
- “The rules align with the Cybersecurity and Infrastructure Security Agency’s July binding operational directive on “prioritizing security updates based on risk.”
- “FedRAMP is directing all certified cloud services to migrate to the new rules by Dec. 7, but will allow vendors to submit corrective action plans for any requirement they can’t meet by that date. The program is giving vendors a three-month “grace period” through March 7, 2027, to meet the requirements.”
- Cyberscoop tells us,
- “Police in nine countries have dismantled large parts of the criminal infrastructure behind three malware kits that enabled ransomware attacks and data theft, seizing 235 servers and 257 domains and identifying credentials belonging to 6.4 million victims, Europol announced Thursday [November 12, 2026].
- The operation, carried out with private-sector partners, targeted the “cybercrime-as-a-service” market, where malware toolkits are built and maintained by one set of actors and sold or rented to others who use them for intrusions, extortion and fraud. Agencies from Bulgaria, Canada, Denmark, Germany, Moldova, the Netherlands, Poland, the United Kingdom and the United States took part, alongside Europol and Eurojust, as part of Operation Endgame, which Europol calls the largest international effort against ransomware enablers, with more than 30 public and private partners.
- “Cybercrime knows no borders and this case is an excellent example of international partnerships working to protect people around the world,” said U.S. Attorney Peter McNeilly for the District of Colorado. “I am proud of the role we played in disrupting this botnet and thank our partners for their exceptional work to unearth this complex network.”
- and
- “The Justice Department on Thursday [November 12, 2026] announced coordinated actions against North Korean groups’ campaigns that have stolen billions of dollars in cryptocurrency, seizing cloud infrastructure behind a social engineering scheme known as FaceStealer and the court-authorized seizure of more than $1.2 billion stolen from cryptocurrency exchange Binance.
- The larger of the two operations is tied to the theft of more than $6 billion in virtual assets from Binance, which the FBI has attributed to North Korean military hackers it calls TraderTraitor. Investigators working with private virtual asset service providers have frozen more than $2.7 billion of the stolen funds, and courts have authorized the seizure of more than $1.2 billion of that amount.
- A. Tysen Duva, the assistant attorney general in charge of the DOJ’s Criminal Division, said its computer crime section and partners froze the stolen Binance assets “before they got beyond law enforcement reach,” halting “about half of this historic breach.”
- The FBI also issued an alert Thursday that detailed the scheme’s mechanics. Hackers known publicly as Sapphire Sleet, BlueNoroff, APT38 and Stardust Chollima hijack accounts on Telegram, LinkedIn, Slack and Discord and invite targets to video meetings. Links lead to typo-squatted domains, hosting imposter versions of Zoom and Microsoft Teams “designed by the hackers to be indistinguishable from the genuine interfaces.”
- Dark Reading adds,
- “Attackers in Binance Breach Targeted Known Weakness, Missing Controls
- “Blockchain complexities, such as attacker-deployed sweepers, hobble security controls designed to limit the damage of an exchange compromise.”
- “Attackers in Binance Breach Targeted Known Weakness, Missing Controls
From the cybersecurity breaches and vulnerabilities front,
- The Wall Street Journal reports,
- “The next AI cyber threat isn’t after your data. It’s after your devices.
- “Connecting everything to the internet has made equipment from printers and ovens to cars and power plants more convenient to run. It’s also made them potentially more susceptible to hackers. Now, following cyberattacks on supermarket freezers and blood banks across Europe, security experts are warning of a sharp rise in such incidents, supercharged by AI.”
- and
- “Attacks on physical equipment used to require specialist skills, time and money. AI is removing those barriers, which means far more connected devices are becoming viable targets,” said Gianni Cuozzo, chief executive of Exein, a Rome-based company that develops cybersecurity technology to protect connected devices, physical AI systems and critical infrastructure.
- “Since last month, supermarket-refrigeration controllers across Europe have been hacked to defrost and heat up units, turning them into ovens and ruining stock. French blood banks were also attacked, though the hack was thwarted before causing damage. The U.S. and France suspect that Iran was behind the cyberattack on European supermarket refrigerators and freezers.
- “Automated physical systems such as electricity grids and controls in water facilities “are not built on the strong foundations they require,” said Richard Horne, the CEO of Britain’s National Cyber Security Centre, in his annual review speech last week.
- “Our resulting exposure as a society is huge and increasingly exploited in cyberattacks, as we are seeing in the disruption of refrigeration units across Europe,” he said, warning of an unprecedented surge looming in cyber incidents.”
- and
- “One new strategy for safeguarding hardware focuses on embedding defenses directly into devices via microchips, rather than relying solely on software or network firewalls. For example, a refrigerator could identify abnormal system behavior and reject a network command to heat to a dangerously high temperature.”
- Cybersecurity Dive adds,
- “A suspected China-nexus threat group used an AI pen-testing framework in a campaign targeting key U.S. sectors, including advanced semiconductors, U.S. defense industry and civil-society organizations, according to a report released Thursday by Palo Alto Networks.
- “The attackers combined the Cairn framework with two Chinese open-source models, DeepSeek and Zhipu GLM, to launch attacks against at least 19 organizations, according to researchers.
- “The attackers successfully gained access or exfiltrated data from 10 of the targeted organizations, according to the report. The hackers obtained dozens of valid credentials, created administrator accounts inside a production system and exposed more than a million records of personal data at one organization.
- “Combining the Cairn tooling with the open-source models allowed the hackers to scale the attacks well beyond normal hands-on-keyboard behavior.”
- and
- “ConnectWise said it is investigating reports that a threat actor is manipulating third-party extensions in ScreenConnect environments to execute malicious remote commands against end users.
- “In a precautionary measure, ConnectWise has disabled access to RMM+ extensions at about 400 environments. The company advises on-premises partners to disable their third-party extensions because it cannot take unilateral action in those instances.
- “While the initial access method remains unconfirmed, our initial assessment suggests the activity may involve a threat actor’s abuse of independent third-party extensions installed in partner environments,” the company said in a statement.”
- Beckers Health IT relates,
- “A data breach involving healthcare AI company Xsolis has ensnared at least 10 health systems and 1.4 million patients.
- “Xsolis, which provides utilization management and care coordination technology for health systems, reported the targeted phishing attack June 5, noting that on Jan. 20 an unauthorized person accessed portions of the company’s IT environment and acquired a “limited number” of files. The company told HHS on June 5 that 1.4 million individuals were affected by the breach, a number the agency posted June 22.
- “Xsolis said it is not aware of any misuse of data and is notifying affected patients. A spokesperson told Becker’s the company isn’t commenting beyond its June 5 statement on the matter. Legal action over the breach has already been filed.”
- Bleeping Computer adds,
- “The Baker Tilly Advisory Group is notifying clients that a data breach has compromised certain personal information about them and their dependents.
- “The company said the exposed data was stored on a platform operated by a third-party service provider Baker Tilly uses to serve clients.
- “Baker Tilly learned of the incident on September 6 and completed its review of the affected data on October 28, when it identified whose information was involved.” * * *
- “Data from the Office of the Texas Attorney General reports that Baker Tilly impacted 10,456 individuals in the state of Texas alone. The portal lists Social Security numbers (SSNs), full names, addresses, and other information as exposed data types.
- “Baker Tilly is a major accounting firm in the U.S., providing auditing, tax planning, compliance, advisory, and business operations services. Globally, the company operates in 147 countries and reports an annual revenue of $6.8 billion.”
- CISA added twelve known exploited vulnerabilities to it catalog this week,
- November 9, 2026
- CVE-2026-102895 GPUStack Improper Authentication Vulnerability
- CVE-2026-116779 SignalWire FreeSwitch Code Injection Vulnerability
- CVE-2026-123158 OpenBSD OpenSSH Unexpected Status Code or Return Value Vulnerability
- CVE-2026-123218 IBM Aspera Faspex Incorrect Authorization Vulnerability
- November 10, 2026
- CVE-2026-20320 Cisco BroadWorks Improper Restriction of XML External Entity Reference Vulnerability
- CVE-2026-111560 Barracuda Networks Multiple Products OS Command Injection Vulnerability
- CVE-2026-111793 Openwave Messaging Limited Webtop Media Transcoder Incorrect Resource Transfer Between Spheres Vulnerability
- CVE-2026-116606 Barracuda Networks Email Security Gateway (ESG) Authentication Bypass Using an Alternate Path or Channel Vulnerability
- CVE-2026-124194 Microsoft Office Outlook Missing User Interaction for Potentially Unsafe Action Vulnerability
- CVE-2026-125350 Google Chromium Use-After-Free Vulnerability
- November 13, 2026
- CVE-2026-114713 UnrealIRCd Improper Neutralization of Delimiters Vulnerability
- CVE-2026-123039 Check Point Multiple Products Incorrect Permission Assignment for Critical Resource Vulnerability
- November 9, 2026
- Cyber Insider points out,
- “Binance warned that an iOS sandbox escape called worse_query could allow malicious apps to read files from other apps, potentially exposing cryptocurrency wallet seed phrases, private keys, and login credentials.
- “Binance operates a cryptocurrency exchange, while self-custody wallets let users control funds through their own cryptographic keys. Exposing a wallet’s seed phrase or private key can let someone else take control of its assets, making access to locally stored secrets particularly consequential.
- “In its warning published today [November 10, 2026], the cryptocurrency exchange identified iOS 26.7.1, 27.0.1, and 27.2 as affected and urged users to remove apps installed from untrusted sources. It stressed that the problem concerns iOS itself and extends beyond Binance.
- “The public proof of concept, published by researcher forcequitOS, describes a broader range of supported iOS versions, from 26.0 through 26.7.2, alongside iOS 27.2. The researcher calls worse_query the successor to an earlier technique named bad_query and describes the standalone code as a developer proof of concept.”
- Bleeping Computer adds,
- “Security researcher Abdelhamid Naceri (also known as Nightmare Eclipse) has released a new Windows zero-day exploit that can let attackers gain SYSTEM privileges.
- “Named SignDrop, it exploits mitigations Microsoft introduced in October 2025 to defend against authentication attacks targeting vulnerabilities in the Common Log File System (CLFS) driver. Naceri added that attacks exploiting CLFS vulnerabilities dropped sharply after the mitigation, but that Microsoft’s mitigation itself introduced this new vulnerability.
- “According to Naceri, the flaw is in how CLFS handles creating the authentication file (“.cnpf”) when it creates a new log file. He said that CLFS enforces access checks correctly at the start, but the logic can be abused if a different, privileged process later opens the log file.”
From the ransomware front,
- Cyberscoop reports,
- “A ransomware attack discovered Monday [November 9, 2026] at Vanderbilt University in Nashville, Tennessee appears to have compromised vast amounts of student and employee personal data, prompting the university to take several digital services intermittently offline as it continues responding to the incident.
- “As of Tuesday morning, the university shared an update that select campus-wide systems and applications were safe to use again, including campus WiFi, Okta, Outlook, Brightspace, Zoom and YES, while others remain restricted, including CampusPrint. Students, faculty and staff were also allowed to resume use of Vanderbilt devices, following orders to shut them down Monday.
- “Vanderbilt Information Technology continues to work with government agencies and other partners to manage this situation, and we will continue to send updates today and throughout this week for as long as necessary,” the update said.
- “A ransomware group known as BYOD has taken responsibility for the attack, according to a post on the group’s leak blog. The group, which emerged in September, has since taken credit for several attacks targeting academic and health care institutions including one last month on the University of New Mexico Health Sciences Center that caused system outages, disrupted patient care and impacted academics.”
- Dark Reading relates,
- “A data extortion outfit has returned to previously targeted law firms for a second go, via a “vishing”-based social engineering campaign where they purport to be IT staff.
- “That’s according to the ReliaQuest Threat Research team, which today published a blog post disclosing that Luna Moth, a threat extortion group that has targeted law firms since at least mid-2023, has for now abandoned remote-management tools and manual exfiltration to steal data, instead looking to swindle victims out of access tokens. Then, they’re using these tokens to access and steal information.
- “Joseph Keyes, author of the blog post and cyber threat intelligence analyst at ReliaQuest, wrote that this latest wave of targets are “likely” those that didn’t pay Luna Moth’s first extortion demand, with this second hit coming weeks-to-months later.
- “A ReliaQuest spokesperson tells Dark Reading that these firms were targeted a second time because the group knows these organizations can be socially engineered (so it tries again with different tactics), for a second set of data. Moreover, as fewer organizations are paying, law firms make a good fit to target, because “their client data is highly sensitive and the reputational hit from exposure is severe, so repeated leverage is worth the effort.”
- Baracuda tells us,
- “Dire Wolf has claimed victims across multiple countries and industries, but its operators, lineage, typical initial-access methods, and organizational structure remain largely unknown.
- “Researchers understand the ransomware better than the intrusion that precedes it. A single documented incident provides the clearest view of early-stage activity, but there is no established playbook for this group.
- “Dire Wolf ransomware aggressively prepares systems for encryption and makes recovery harder. Technical analyses document process and service termination, backup and recovery sabotage, partial file encryption, the .direwolf extension, and victim-specific extortion through the HowToRecoveryFiles.txt ransom note.”
- Recorded Future informs us,
- “GhostFrame is a previously undocumented post-exploitation framework that gives threat actors extensive control over compromised VMware ESXi environments. Once an attacker gains root-level access to an ESXi host, GhostFrame provides a single interface to access virtual machines running on the hypervisor, harvest credentials, execute commands, and pivot deeper into victim networks. Insikt Group identified GhostFrame being used by TAG-218, a threat activity group whose operations overlap with Settra ransomware.
- “GhostFrame is notable because it allows threat actors to operate from the virtualization layer, where a single compromised ESXi host can provide access to multiple virtual machines and the systems and data they contain. Insikt Group’s findings demonstrate how attackers can use compromised hypervisors as a centralized platform for post-exploitation activity, potentially expanding access across victim environments without compromising each virtual machine individually. The discovery highlights the importance of treating virtualization infrastructure as a high-value target and monitoring ESXi hosts for signs of unauthorized access and post-exploitation activity.”
From the cybersecurity business and defenses front,
- Cyberscoop reports,
- “Policymakers from Congress and the executive branch are looking for ways to further boost the U.S. quantum computing industry, including by growing industrial capacity to translate scientific breakthroughs into workable technology.
- “Deputy Secretary of Commerce Paul Dabbar said the American industry is “by far the greatest leader in quantum technologies,” with a tech stack that is the envy of the world. The challenge will be translating those technological gains into real, brick-and-mortar industrialization and manufacturing dominance.
- “Dabbar pointed to other examples of federal support, such as grant funding from the CHIPS and Science Act and letters of intent signed earlier this year with nine companies investing more than $2 billion into different quantum technologies.
- “But one of the primary questions that Commerce officials ask companies when making decisions on federal investments is “How do we move a bit further down the road from discovery towards actually supporting individual technologies that can commercialize?” he said.”
- and
- “The United States this week signed on to an international set of industry standards for commercial offensive cyber tools that human rights groups and other advocates wish had been stronger.
- “The Pall Mall Process voluntary industry guidelines for commercial cyber intrusion capabilities (CCICs), such as spyware, are the second step in a U.K.- and France-led initiative, following last year’s code for nations.
- “Through the Industry Guidelines, we aim to disincentivise irresponsible activity across the market for CCICs; improve transparency for regulators, vendors and customers; and provide clarity for businesses operating both within and peripherally to the industry,” the U.K. and French foreign ministries said after the guidelines were finalized Wednesday.
- “The guidelines have four main pillars: compliance with regulations and norms, due diligence, risk management and monitoring of compliance. They contain recommendations like complying with export controls, creating a risk assessment process, implementing technical controls on access such as encryption and continuously observing who can use the capabilities.”
- Computer World relates,
- “Anthropic has introduced Claude Fable 5.2 and Mythos 5.2, claiming significant improvements in handling complex software engineering tasks while cutting API prices by 20% compared with the previous versions.
- “Fable 5.2 and Mythos 5.2 share the same underlying model but have different safeguards. Fable is broadly available, while direct access to Mythos is restricted to vetted cybersecurity and life sciences organizations to limit potential misuse.
- “The models are both priced at $8 per million input tokens and $40 per million output tokens, down from $10 and $50, respectively. Anthropic said the newer versions can generate output up to 30% faster than their predecessors.”
- The Register adds,
- “Anthropic has started extending access to its bug-finding Mythos model by allowing Cisco and IBM to make it part of their security products.
- “Cisco announced its integration today at an event in Melbourne, Australia, where it revealed that the red teaming tools in its AI Defense suite, which looks for weaknesses and inappropriate activity in AI models, agents, and applications they power, will be able to call Mythos to test customers’ setups.
- “At a very high level, AI Defense has the ability to look at models and break them,” said Jeff Schultz, Cisco’s senior vice president for portfolio strategy and marketing. He said the tool is now “super-charged by Mythos” and will identify more danger that AI Defense will then address.
- “AI Defense already uses AI-powered red teaming to find flaws. Mythos will become another option. Cisco will make it available under users’ existing licenses, without additional cost.”
- The Cisco deal exposes the functionality of Mythos to end-users, without letting them use the model itself.
- * * * The Register asked [Anthropic’s Natalie] Mead if Anthropic has similar deals with other third parties and she demurred. That may be because a few hours after Cisco told us about its use of Mythos, IBM announced that it will also integrate Anthropic’s model into a product – namely its Concert tool, which uses AI to scan applications for risks and weaknesses.
- IBM says it will use Mythos “to uncover weaknesses traditional scanners miss, determine which pose real risk and generate fixes for developer approval.”
- SC Media tells us,
- “It’s no secret that artificial intelligence is transforming cybersecurity by accelerating the speed and rate of attacks, expanding enterprise attack surfaces and giving autonomous systems access to sensitive resources, whether or not they need it.
- “According to Akamai’s latest State of the Internet report, “Speed, Scale, and Nonhuman Identity: The Agentic Threat Landscape,” organizations must reassess the readiness of defenses that were built around human response times and conventional security boundaries.
- “AI-powered attacks can exploit vulnerabilities more quickly than security teams can patch them. Meanwhile, AI agents, browser extensions and interconnected applications can all introduce new opportunities for data theft, unauthorized access and manipulation. * * *
- “Akamai identifies four priorities for securing the emerging agentic enterprise.
- “Shift to adaptive edge governance. Deploy runtime defenses, API filters and isolation mechanisms capable of blocking emerging threats before patches become available.
- “Lock down the browser edge. Monitor AI extensions, enforce appropriate permissions and prevent sensitive information from flowing into unauthorized AI services.
- “Protect brand authority in a zero-click economy. Monitor how AI crawlers access and represent corporate information, addressing inaccurate citations, misleading responses and reputational manipulation.
- “Match autonomy to verifiability. Grant AI agents authority according to how reliably their actions can be checked and how reversible potential mistakes are. Require human oversight for consequential decisions.”
- and
- “AI agents are now a disclosed business risk. Apple’s latest annual report warns that AI agents can take actions autonomously, including in systems they may not be authorized to access, and that some consequences, such as deleting data or completing transactions, may be difficult or impossible to reverse.
- “Apple then added that it may have limited ability to oversee and control third-party artificial intelligence agents and other technologies.” * * *
- “If we can’t fully test an agent, can’t control how it was built, and can’t reverse what it does, the control has to sit at the one point we do own: the moment the agent acts inside our environment.
- “In practice that comes down to five disciplines:
- “Know what’s running: Keep a live inventory of every agent, on laptops, in pipelines and inside SaaS platforms, along with what each one can reach and whose credentials it uses. We can’t govern an agent we haven’t found.
- “Scope authority to the task: An agent that summarizes tickets doesn’t need write access to production. Grant access for the job at hand, and treat an agent that inherits everything its user can do as a finding.
- “Gate the irreversible: The ability to delete data, move money, change permissions, and send information outside the company should require an explicit approval, or get blocked outright. A good place to start: Apple’s own example list, deleting data and completing transactions.
- “Keep evidence that doesn’t depend on the agent: Apple’s filing names concealment as a risk. Our record of what an agent did should come from where the action happened, tied to who launched it and what it was asked to do, and never from the agent’s own account of its work.
- “Test agents the way attackers will: Do this with poisoned documents, hostile web pages, and ambiguous instructions, before an incident does it on its own.”
- Here’s a link to Dark Reading’s CISCO Corner.
