The security & privacy stories our desk judged worth your time, written so anyone can follow them.
Infisical raised $40 million to help companies let AI agents work without ever seeing real passwords, a risk for any business that connects AI to its accounts.
Read more: Infisical raises $40M to keep AI agents away from company passwordsA cyberattack at UNM Health paused surgeries and diverted ambulances. Services are back, but patients should expect scam messages that use their care details.
Read more: Ransomware hits New Mexico's biggest hospital system, diverting ambulancesA federal appeals court found the Fed let banks count costs the law did not allow in its 21-cent debit fee cap, which could change what stores pay.
Read more: Court says the Fed's 21-cent debit card fee cap was set too highA shared analytics script at Convertr, a Brazilian checkout provider, was altered to copy card details. Shoppers who paid there since October 24 should act.
Read more: Hackers slip card-stealing code into checkout pages of Brazilian online storesSecurity firm Cathedral says its AI agents got into Anthropic's business systems through connected apps. Anthropic confirms limited access but says Claude was not reached.
Read more: Researchers say AI tools got into Anthropic's business systemsAt an industry summit, payment leaders said OTPs and PINs no longer stop scams where victims send the money themselves. New checks aim to catch trouble first.
Read more: India's payment apps are learning to spot scams before money movesA Thetford surgery says an automated call about 'updating your GP records' is not from its team. Hang up and check before sharing any details.
Read more: Surgery warns patients about robocalls claiming to update GP recordsResearchers found attackers in Japan using AI tools to automate break-ins, which means weak passwords and unpatched devices get found faster than ever.
Read more: Hackers in Japan Are Using Teams of AI Helpers to Break Into NetworksMicrosoft released emergency Windows updates on October 26 to block certificates tied to eight weak eMudhra keys that could be used to fake website identities.
Read more: Microsoft pulls trust from weak digital IDs on Windows computersPaylogix, a New York benefits firm, says hackers copied files on 41,739 people, including Social Security numbers. Three other health groups also reported breaches.
Read more: Hackers Took Data on 41,739 People From a Benefits FirmAnthropic says a China-based studio ran 20-plus dating apps with AI personas and human video callers, so a warm chat or a video call no longer proves a real person.
Read more: One studio used AI to run 4,700 fake dating personasThe Better Business Bureau says scammers are phoning people during open enrollment, posing as agents who promise a cheaper health plan and ask for private numbers.
Read more: Fake Medicare helpers are calling during open enrollmentMastercard is bringing a tool to Europe that helps banks tell real purchases from fraud, which could protect shops from losing money on genuine sales.
Read more: Mastercard expands tool to curb false disputes over real purchasesOverwatch Data raised $15 million for AI tools that help banks and payment firms catch attackers who now work faster and cheaper than ever.
Read more: Overwatch Data raises $15 million to help firms fight AI-powered hackersSaaStr's AI agents destroyed data the business needs to run, twice in a few weeks. Both times, an unplanned copy saved it. Your business could be next.
Read more: Two AI helpers wiped key business data, and an accidental copy saved itSecurity firm Surfshark says a fake TikTok Shop site caused nearly one in five phishing blocks and is still live, just before Black Friday shopping.
Read more: A fake TikTok Shop is behind one in five phishing blocksLloyds now uses AI helpers to guide staff through debit card fraud cases, and says the tool's decisions are right more than 90 per cent of the time.
Read more: Lloyds bank now uses AI helpers to investigate debit card fraudA new report says shoppers now use card disputes to complain about late packages and unwanted subscriptions, while fraud claims drop to a two-year low.
Read more: More Shoppers Dispute Card Charges Over Late Orders, Not FraudHuntress says attackers now hide in remote IT tools and Microsoft 365 logins, so a break-in can look like normal work and set off no alarms.
Read more: Hackers Are Hiding Inside the Work Tools Your Business Already TrustsScammers are posing as Madison County, Florida, zoning staff and demanding money by call, text and email. The county says real staff never do that.
Read more: Fake county officials are threatening property owners for moneyPolice warn of fake firecracker sales, parcel links and bogus bookings before Diwali. Nationwide online fraud losses topped ₹22,495 crore in 2025.
Read more: Festive scam season: fake sales and parcel links ahead of DiwaliHackers say they have posted data from GMO Internet's Onamae cloud desktop service, used by traders. Customers are being told to reset passwords.
Read more: Hackers post data allegedly stolen from Japanese web giant GMOFord says its $28,000 Fathom electric pickup has the power converter built in, which could make home blackout backup cheaper to set up.
Read more: Ford's $28,000 electric truck can power your house in a blackoutA survey finds older Americans are far more likely than young adults to say a person tricked by a scammer should absorb the loss, not the bank.
Read more: Older Americans more likely to blame scam victims for their losses
