EE Scam Guard launches AI Triple-Lock to fight UK mobile fraud

None of this replaces basic hygiene. The Financial Conduct Authority’s ScamSmart guidance and the Information Commissioner’s Office published advice on nuisance and scam communications are the regulator-side reading that any UK mobile customer should bookmark. Our own roundup of the best password managers for 2026 covers the credential layer most phishing campaigns are ultimately trying to harvest, and the hardware security key argument applies just as cleanly to a banking app as it does to an AI account.

Where EE Scam Guard realistically helps, and where it doesn’t

Be specific about the wins. Network-level call labelling, the SMS layer of Triple-Lock, and email scanning in supported mail apps are genuinely useful against the largest UK fraud volume: cold scam calls and bulk SMS impersonating delivery firms, HMRC and banks. EE bundling Cifas identity monitoring and breach alerts is the part of the package that arguably justifies the £2 on its own. The reuse of Norton’s Genie engine, rather than a bespoke EE model, is sensible: Norton sees a much bigger global telemetry feed than any single UK telco could build alone.

Now the gaps, because they matter. Scam Guard, by EE’s own description, scans SMS, email and the open web. It does not, and cannot, read end-to-end encrypted chats inside WhatsApp, Signal, iMessage with iMessage-to-iMessage, encrypted RCS, Telegram secret chats, Instagram DMs or Messenger encrypted threads. That is precisely where modern UK scam traffic is migrating, especially “Hi Mum” family-impersonation fraud on WhatsApp, romance and crypto scams on Instagram and Telegram, and marketplace fraud on Facebook Marketplace chat. Any UK operator selling AI scam protection in 2026 needs to be transparent that the most dangerous attack surface sits inside apps the network simply cannot inspect.

The second honest caveat: AI filtering is a probability game. Norton’s Genie will catch a high share of obvious phishing, but advanced AI-generated scam emails and voice clones are explicitly designed to slip past statistical classifiers. EE’s own Cubitt quote (above) concedes AI is “making scams more convincing and harder to detect”. That is not a flaw in Scam Guard; it is a reality of the cat-and-mouse, and any UK reader thinking a £2 add-on means they can stop reading messages carefully is misreading the product.

What UK buyers should actually do this week

Start with the free layer. Forward suspicious texts to 7726 on any UK network. Report scam calls to the same code. Use the NCSC’s Suspicious Email Reporting Service (forward to [email protected]) for phishing emails. None of this costs anything and all of it feeds the data set EE, O2 and VodafoneThree’s filters are drawing from. Our guide to privacy-first mobile workflows walks through the wider data-minimisation routine that makes most of these scams harder to land in the first place.

Then assess the paid layer on its own terms. £2 per month is £24 a year, materially less than standalone Norton 360 or LifeLock-style identity products, and Cifas is the genuine extra. If a household already pays for antivirus or a password manager with dark-web monitoring, the marginal benefit shrinks; if not, EE Scam Guard plus 7726 reporting plus careful behaviour inside encrypted apps is a credible setup, not a substitute for staying suspicious.