SIGNALS: Week 47 - Check, Please

18 November 2026 9 min read Andy

SIGNALS: Week 47 - Check, Please

18 November 2026 | About an 8-minute read

Almost every story this week turns on one small act: the check. The pause before paying, the glance at a sender address, the question asked over a second channel. People believe in checks, recommend them to others, and then skip them at the exact moment they matter. The good news is that several of this week's items, from a national fraud plan in Belgium to a clever lab study from Bristol and Bath, point at the same practical fix: stop relying on willpower and start building the check into the moment itself.

Everyone believes in scams, right up to the checkout

With Black Friday less than a fortnight away, Barclays has published research that captures the awareness-behaviour gap in one tidy dataset. In an Opinium poll of 2,000 UK adults carried out in October, 76 per cent said AI has made online scams more convincing, and around a third had either been targeted by a shopping scam or knew someone who had. Yet only 36 per cent said they would check whether a website is genuine before buying, 21 per cent admitted to shopping online while doing something else, and just one in ten felt confident they could avoid being deceived over the Black Friday weekend.

The bank's own claims data adds a sharper edge. The value of Gen Z purchase scam claims rose 56 per cent in the first half of 2026, against a 14 per cent rise overall, and 46 per cent of Gen Z shoppers said fear of missing out pushes them to buy sale items quickly. Barclays' fraud expert Kirsty Adams put her finger on the mechanism: a countdown timer or a deal from a familiar-looking account gets people to pay before they check.

What does this mean for me

This is a bank-commissioned survey, so treat the headline numbers lightly, but the pattern is one you can act on this week. Seasonal awareness comms usually aim at knowledge, and knowledge is clearly not the problem here. Aim instead at the moment of purchase: short, well-timed nudges about countdown timers, FOMO and distracted checkout behaviour, ideally landing in the days before 27 November. The same psychology applies at work. Finance and procurement teams face their own version of urgency-driven payment pressure, and a Black Friday campaign is a disarming way to open that conversation.

Source: survey commissioned by Barclays, conducted by Opinium, reported by The Independent, 13 November 2026. Fraud experts explain how to avoid AI scams this Black Friday

The other people on the video call may be recordings

The FBI issued a warning on 12 November about a North Korean operation it calls FaceStealer, which has taken more than 400 million dollars in cryptocurrency over 18 months. The method deserves attention well beyond the crypto world. Attackers take over the messaging accounts of earlier victims, read their chat history, and invite the victims' real contacts to a catch-up call. The meeting link, sent shortly before the call starts, points to a lookalike Zoom or Teams domain. On the call, the target sees and hears their actual contact because the attackers play recordings captured from that contact's webcam during an earlier attack. Then a chat message appears: " Others cannot hear you; run this quick fix. The "fix" installs malware; the webcam has been recording since the link was clicked, and the new victim's face becomes part of the next call.

The FBI has seized more than 900 such recordings. Notably, it reports that stolen footage of real people still makes up most of the scheme, with deepfakes only beginning to appear.

What does this mean for me

Most deepfake awareness advice tells people to look for glitches, odd blinking, mismatched lighting. This scheme sails straight past all of it, because the footage is real. The lesson to teach is verification behaviour, not detection skill: join meetings through the official app with an invite code rather than a link, confirm unexpected invitations through a second channel, and treat any request to run a fix or paste a command as a stop signal, full stop. That last behaviour also covers the ClickFix-style lures that have dominated the last few months. Anyone with access to a code repository, a wallet, or production credentials should hear this story in full, because it is memorable in a way generic deepfake warnings aren't.

Source: FBI IC3 public service announcement, 12 November 2026, reported by Help Net Security. Stolen webcam clips of 900+ victims feed North Korea's $400 million fake Zoom scheme

Belgium decides phishing is a design problem

On 12 November, Belgium's consumer protection and interior ministers launched a National Anti-Online Fraud Plan containing 29 measures, developed with banks, telecoms operators, platforms and police. Belgians reported more than 11,000 phishing cases in 2025, up 30 per cent in a year, losing around 93 million euros. What makes the plan interesting for this audience is where it places the burden. Phish Nemo, a police project that shows a warning page in front of known fraudulent websites, is credited with heading off an estimated 100,000 potential victims and will be rolled out nationwide. A new platform called TILDA lets banks share intelligence on suspicious accounts so that a fraudster closed down at one bank cannot simply open an account at another. The itsme identity app will compare the user's face with their eID photo for certain actions, a public reporting form on Safeonweb will feed verified reports to telecoms firms and platforms for faster takedowns, and police and bank staff will be trained to support victims more empathetically.

What does this mean for me

Here is a government putting prevention into the infrastructure rather than asking citizens to try harder, and the ministers claim it is the first plan of its kind in the EU. It is a useful benchmark for anyone making the case internally that human risk is reduced by system design as much as by education. The translation exercise is worth running with your own organisation: what is your equivalent of the warning page at the point of click, the shared intelligence between teams, the low-friction reporting route, and, often forgotten, the empathetic treatment of people who have already been caught?

Source: Belgian government announcement, reported by VRT News, 12 November 2026. How does the Belgian government plan to tackle online fraud?

A good habit only works if you know what you are looking for

A peer-reviewed study in the November issue of Computers and Security, by Tobias Weickert, Adam Joinson and Barnaby Craggs of the Universities of Bath and Bristol, tested whether secure habits actually protect people from phishing. The team built PhiT, a simulated email client developed with the UK's National Protective Security Authority, and gave participants an in-tray exercise: work through about 20 emails against the clock, six of them phishing, in one of three job roles. Across 88 usable sessions, the headline finding is usefully uncomfortable. A reliable habit of checking sender addresses was associated with falling for fewer phishing emails, but only in the IT specialist scenario, where participants plausibly knew what a suspicious sender actually looks like. Elsewhere, the habit made no measurable difference. The authors' conclusion: checks protect only when they are grounded in knowledge of deception cues, and they propose "cue engineering", building prompts such as new-sender warnings into the interface so the habit has something solid to anchor to.

What does this mean for me

"Always check the sender" is probably the most repeated line in security awareness, and this study suggests it is half an instruction. Pair the habit prompt with the diagnostic knowledge: show real lookalike addresses from your own sector, not generic examples. And work with what your email client already offers: external sender banners, first-contact flags, unverified sender warnings, because those are cue engineering in practice. The sample is small and the design exploratory, so keep the specifics loose, but the direction of travel fits much of what this field has learned about habits.

Source: peer-reviewed, open access study, Computers and Security, November 2026 issue. Can secure habits counter phishing? An exploration using a novel in-tray simulation

Untested confidence at the plant gate

The SANS 2026 ICS/OT survey, drawing on 696 industrial practitioners plus a smaller leadership sample, contains one of the starkest perception gaps of the year. Asked whether their plants would keep running through a full IT compromise, every single leader said yes: 38 per cent expect to run indefinitely on islanded systems, 62 per cent to get by temporarily on manual controls. Among the practitioners who would actually operate those manual controls, 15 per cent said critical processes could sustain safe production in manual mode beyond 48 hours. Forty per cent did not know. Meanwhile, 34 per cent of organisations had an ICS/OT incident in the past year, 1.7 times last year's rate, and 68 per cent said losing just one or two veteran employees would delay response or create severe operational risk. Report author Jason Christopher's line is worth stealing: an untested assumption is not a capability.

What does this mean for me

Even if you never go near a plant, this is a clean illustration of how confidence concentrates at the top while capability, and doubt, live at the front line. The fixes are behavioural: only 58 per cent of tabletop exercises include frontline engineers and operators, so put the people who hold the manual knowledge in the room, and treat that 68 per cent veteran-dependency figure as a human risk register item. Asking "who actually knows how to do this by hand, and when did we last watch them try?" is a question worth exporting to every function, not just operations.

Source: SANS Institute 2026 ICS/OT survey, reported by Help Net Security, 12 November 2026. CISOs say OT will survive a cyberattack, but their engineers aren't so sure


That's the week. Five stories, one refrain: a check that lives in the system, or in a well-anchored habit, beats a check that lives in good intentions. See you next Monday.

The CyBehave Team

← All Signals briefings

Get Signals in your inbox

Subscribe and choose the Signals weekly briefing to receive each edition the morning it publishes.