The Department of Know: Sedgewick confirms incident, Coupang store credit only, AI needs generators

This week’s Department of Know is hosted by Rich Stroffolino with guests Peter Clay, CISO, Aireon, and Chris Ray, Field CTO, GigaOm

Missed the live show? Check it out on YouTube

The Department of Know goes live every Monday at 4:00 p.m. ET. Join us each week by registering for the open discussion at CISOSeries.com

Palo Alto Networks boss calls AI agents biggest insider threat

This is according to Chief Security Intel Officer Wendi Whitmore, describing them as the new insider threat to companies in 2026. She describes the threat as a double edged sword, because “CISO and security teams find themselves under a lot of pressure to deploy new technology as quickly as possible, and that creates this massive amount of pressure to go through procurement processes, security checks, and understand if the new AI applications are secure enough for the use cases that these organizations have.” She added “One of the risks stems from the superuser problem, which occurs “when the autonomous agents are granted broad permissions, creating a superuser that can chain together access to sensitive applications and resources without security teams’ knowledge or approval. More details from her interview are available in the show notes to this episode.

(The Register)

Will prompt injection ever be ‘solved’?  

OpenAI says prompt injection attacks against browser-based AI agents like ChatGPT Atlas may never be fully eliminated, after internal red-teaming uncovered a new class of attacks that can hijack agents during routine web workflows. The company shipped an update with a newly adversarially trained model and stronger safeguards, but warned that agents with access to email, documents, and web services are inherently higher-value targets. 

(CyberScoop)

Korean Air shares supplier attack

Korean Air disclosed a data breach after its in flight catering and duty free subsidiary Korean Air Catering and Duty Free, or KC&D, was hacked, exposing personal data tied to about 30,000 airline employees. The airline says customer data was not affected and that the leaked information appears limited to employee names and account numbers stored on KC&D’s ERP system. The Clop ransomware group has claimed responsibility for the KC&D attack and says it has already leaked the stolen data. (Security Affairs)

DarkSpectre campaigns exposed

Koi Security researchers say a Chinese-linked threat actor they track as DarkSpectre has run three long-running malicious browser extension campaigns that together impacted more than 8.8 million users across Chrome, Edge, Firefox, and Opera. The campaigns, dubbed ShadyPanda, GhostPoster, and Zoom Stealer, used legitimate-looking extensions to hijack searches, commit ad fraud, and quietly collect sensitive corporate meeting data from platforms like Zoom, Google Meet, and Microsoft Teams. Koi describes the operation as infrastructure for large-scale corporate espionage, rather than consumer fraud. 

(The Hacker News)

Huge thanks to our sponsor, Hoxhunt

A small tip for CISOs: if you’re unsure whether your security training is actually reducing phishing risk, check out what Qualcomm achieved with Hoxhunt. They took their 1,000 highest-risk users from consistent under-performers to outperforming the rest of the company, driving measurable human risk reduction and earning a CSO50 Award. See the Qualcomm case at hoxhunt.com/qualcomm

Diesel generators and aircraft engines in high demand to power AI

The developers of data centers are now using aeroderivative turbines, “based on or made from jet engines,” as well as diesel generators, to address a growing need for power to process AI technology. This is being done to counter the issue of supply chain shortages and wait times of up to seven years to connect to the grid, as well as growing backlash over their impact on consumer utility bills. This power is needed for the training and running of artificial intelligence models. As a result, local and federal regulators in the U.S. are starting to loosen the restrictions on the use of backup generators and are even floating the idea of commandeering existing backup generators, such as those located behind many large stores and businesses, in order to support the demand.

(Financial Times)

Coupang to spend over $1 billion to compensate users

South Korea’s leading e‑commerce and logistics platform, often dubbed “Korea’s Amazon” has announced the amount, specifically the equivalent of $1.17 billion to compensate almost 34 million people affected by its recent data breach. This will be in the form of purchase vouchers for impacted users. Coupang plans to “sequentially notify all 33.7 million account users via text message regarding the voucher redemption process.”

(Security Affairs)

Sedgwick confirms New Years Eve cyber incident

The claims administration company has confirmed that its government-focused subsidiary is dealing with a cybersecurity incident. Sedgwick “provides claims and risk management services to federal agencies like the DHS, Immigration and Customs Enforcement, Customs and Border Protection, Citizenship and Immigration Services, the Department of Labor, and CISA. The TridentLocker ransomware gang has claimed responsibility. Sedgwick stresses that its Government Solutions arm is segmented from the rest of its our business, and that “no wider Sedgwick systems or data were affected.”

(The Record)

Steve Prentice
Author, speaker, expert in the area where people and technology crash into each other, viewed from the organizational psychology perspective. Host of many podcasts, voice actor and narrator for corporate media and audiobooks. Ghost-writer for busy executives.