*make sure to check your spam folder for the signup acceptance email.
|
Interesting week this week, as AI continues to embed itself into our stories, our organisations and our personal lives. Even in insider risk and threat, it feels like each week there is a new story to talk about. There are always different avenues for insider risk. This goes back to what I do day to day, when we run the workshops every six weeks. Each time there's a new, novel way AI can be used for data exfiltration, either by a human or by the agents themselves. I think we're still before the horizon of large-scale breaches, particularly AI being weaponised by criminal gangs or nation-state actors. Headline: The Security Chief Was the Hacker
What a complex case. This is a trusted insider from a cyber investigation standpoint. He was brought in to help investigate a breach, then allegedly took that information and used it to improve the malware he was building. The investigators thought they were handing it to someone on their side. There is also an insider angle in who he targeted: people in the community, reached through these organisations. These types of insiders are potentially the most dangerous ones out there. Source: Israeli cyber manager accused of remotely accessing cameras, stealing passwords and infiltrating 26 companies (CTech by Calcalist) Source: מנהל אבטחת מידע בקבוצת חמת מואשם במאות עבירות סייבר (N12, Hebrew) Insider Scoop: Hackers Claim the FBI's Own Staff List
The insider angle here is the exposed staff in counterintelligence roles. If the data is real, foreign intelligence services could use it to blackmail or leverage them. Having their personal details out there makes them more visible targets, and their families too. The exposure of them and their families can be used to leverage them into insider activity. Source: Exclusive: Hacked FBI data has sensitive information about employees' intelligence roles (Reuters via Yahoo News) Source: Hacking group ShinyHunters claims it breached the FBI, stole agents' and applicants' data (TechCrunch) The Politicians and the Pandemic MoneyLawrence mayor Brian DePeña
Source: Massachusetts Mayor Indicted by Federal Grand Jury for Pandemic Loan Fraud and Money Laundering (US Attorney, D. Mass.) Source: Lawrence mayor hit with second indictment this summer, now over alleged wiretapping (WBUR) Massachusetts state representative Francisco Paulino
Source: Massachusetts State Representative Charged With Pandemic Loan Fraud and Money Laundering (US Attorney, D. Mass.) Source: State Rep Charged with Pandemic Loan Fraud and Money Laundering (Massachusetts Inspector General) Again, two stories of people in positions where misuse of funds is involved. We see this week on week. These happen where there is no oversight or checking at the time, and it is only when investigations happen that they are uncovered. DePeña's second case, secretly recording staff and sharing it, plus allegedly intimidating a witness, is potentially another example of malicious insider behaviour. Illegal means used to cover things up or swing situations his way. Crime RoundupThe payroll manager and the fake garnishments.
The contractor whose access outlived the job.
The teller with the old withdrawal slips.
Source: Bank teller accused of stealing $60K from customer accounts, per affidavit (Action News 5) Source: Federal Reserve letter to Demeries Brooks, 12 August 2025 Again, we see people who had access to money, or the management of it, in their jobs. They all went unnoticed for some time while they enriched themselves and stole from their organisation or from others. AI Insider Threat Watch: The Agent That Didn't Take No for an Answer
There's a lot of news and noise around this story at the moment. From an insider risk perspective, if these agents are going off and doing things that are unintended, we should be training them to behave like proper penetration testers rather than insiders. If they do breach a system while trying to find information, then instead of just getting the information and leaving, they should go through a process of responsible disclosure. That way we're not waiting a month for companies like OpenAI to report them to the government or the organisation that was hacked. Source: OpenAI hacked Medicare portal, Prime Minister Anthony Albanese says (ABC News) Source: OpenAI agents 'infiltrated Australian government website' (The Register) Mitigation Corner: Speed Is the Whole Game
The metric to put in front of the board is time-to-contain, not number of alerts. Disclosure: I work for DTEX, and DTEX sponsored this report. Source: 2026 Cost of Insider Risks Global Report (Ponemon Institute and DTEX) Source: Ponemon Cost of Insider Risks: Key Findings 2026 (DTEX) Trusted Profession Watch: The Nurse and the Five Residents
This is another example of a person in a trusted position abusing that position for their own benefit, and potentially hurting those in vulnerable positions. This week is a good opportunity to point to a really interesting podcast called The Retrievals. It covers a very similar situation, but the vulnerable people are women undergoing IVF treatment in the US. Here's the link if you would like to listen to the series. Listen: The Retrievals (Serial Productions and The New York Times, on Spotify) TriviaThe Australian Signals Directorate is part of the taskforce reviewing the OpenAI Medicare breach. What is ASD's motto? A) Always vigilant, never seen B) Know the enemy, guard the gate C) Reveal their secrets, protect our own D) In silence, strength (Answer at the bottom.) UpdatesQuick follow-ups on stories from earlier issues. Correction on the Huawei trial. In Issue #23 we said Huawei pleaded not guilty to 14 counts, with jury selection on 9 September. Four days before trial, prosecutors dropped two sanctions counts, so 12 counts went to the jury. Jury selection began on 8 September, with opening statements on the 9th. Source: United States v. Huawei Technologies Co., Ltd., 1:18-cr-00457 (E.D.N.Y.), docket (CourtListener) ResourcesA few places worth your time if you want to go deeper on insider risk.
Worth bookmarking:
And my own: Insider Threat Awareness Training on YouTube, a free course with certification. Trivia AnswerC, "Reveal their secrets, protect our own." ASD began as the Defence Signals Bureau in Melbourne on 1 April 1947 and took its current name in 2013. This week it is on the other side of that motto, working out how an AI agent got to government files it was never meant to see. |
*make sure to check your spam folder for the signup acceptance email.