Deepfake Video Calls Are Now the Scam: The $25 Million Arup Case

Deepfake Voice and Video Scams Are Surging — and Getting Harder to Detect

AI-enabled deepfake and voice cloning scams have become one of the three most dangerous fraud threats targeting Americans right now. Approximately one in four Americans reports receiving an AI-generated deepfake voice call in the past 12 months. Another 24% say they cannot distinguish deepfake calls from genuine ones.

The scale of escalation is hard to overstate. AI scams spiked 450% in 2025 according to DOJ reporting. Deepfake fraud in North America surged 1,740% between 2022 and 2023. The technology now functions as an amplifier for every other scam type — grandparent emergency scams, IRS impersonation, investment fraud — rather than a standalone threat. AI tools require no technical expertise, cost pennies to deploy, and produce voice clones and deepfake videos that defeat fraud awareness trained on older, more detectable methods.

The broader fraud environment reinforces the urgency. FBI IC3 data for 2024 recorded $16.6 billion in reported cybercrime losses — a 33% increase over 2023. The FTC tallied $12.5 billion in fraud losses specifically, a 25% jump from the prior year. And those figures likely represent a fraction of actual harm: the FTC estimates only 2–6.7% of victimizations are formally reported, putting true losses potentially 15 to 50 times higher.

Gobble's Take: When a quarter of Americans can't tell a cloned voice from a real one, a family code word isn't paranoia — it's basic hygiene.

Source: Vectr-Cast / cyberwarrior76

Toronto Police Seize Mobile Rogue Cell Towers Used to Deliver Scam Texts at Scale

Three men in the Greater Toronto Area were arrested in the spring of 2026 after Toronto Police dismantled what the force described as the first known SMS blaster operation in Canada. The devices — mobile units that impersonate legitimate cellular towers — forced nearby phones onto fake networks and pushed fraudulent text messages directly to those phones, bypassing normal carrier filters entirely.

According to Toronto Police, tens of thousands of phones connected to the rogue devices, and investigators recorded more than 13 million network disruptions. Some of those disruptions lasted from seconds to several minutes and, police said, could have prevented affected phones from reaching 911 during those windows. The devices were not stationary: police described them as mobile, meaning they could move through residential blocks, transit corridors, and event areas, hit thousands of devices, then disappear back into traffic before detection.

Once connected, victims received texts that appeared to come from their bank, Canada Post, the 407 ETR toll highway, or parking authorities. The links led to phishing pages designed to collect banking credentials, payment card details, or identity information. The investigation, called Project Lighthouse, began in November 2025 after a cybersecurity partner alerted police to a blaster operating in downtown Toronto. Dafeng Lin, 27, of Hamilton, and Junmin Shi, 25, of Markham, were arrested March 31, 2026; Weitong Hu, 21, of Markham, turned himself in April 21. Charges include mischief endangering life, personation with intent to gain advantage, and trafficking identity information, among others.

Gobble's Take: A text that looks like it came from your bank — arrived without you clicking anything, triggered just by walking past a parked car — is the kind of scenario worth mentioning to anyone who still trusts sender names in their messages app.

Source: Covert Access Team

FBI Reported $16.6 Billion in Cybercrime Losses in 2024 — AI Is Now Woven Into Nearly Every Scam Type

The numbers reported by federal agencies for 2024 are specific and worth sharing with family members who still think of scams as a niche problem. FBI Internet Crime Complaint Center (IC3) data recorded $16.6 billion in cybercrime losses — a 33% increase over 2023, according to the Vectr-Cast fraud assessment. The FTC's Consumer Sentinel data separately tallied $12.5 billion in fraud losses, a 25% jump from the prior year.

The assessment identifies three categories driving the most harm. Pig-butchering cryptocurrency investment fraud — in which scammers cultivate a relationship over weeks or months before steering victims toward fake trading platforms — was responsible for an estimated $10 billion in annual U.S. losses, according to the report, with the pipeline described as "active and adaptive" despite a $15 billion DOJ forfeiture case in October 2025 and a $578 million cryptocurrency seizure by the Scam Center Strike Force. AI-enabled deepfake and voice cloning scams spiked 450% in 2025 according to DOJ reporting cited in the assessment, and approximately one in four Americans reported receiving an AI-generated deepfake voice call in the past 12 months, with another 24% saying they could not tell whether a call was real or AI-generated. Government impersonation scams — callers posing as IRS agents, Social Security Administration staff, or law enforcement — generated more than 330,000 FTC complaints in 2025, a 25% increase; the IRS added AI-enabled phone impersonation to its 2026 "Dirty Dozen" list of active tax scams, its first time listing that specific threat.

The assessment's framing of AI as an "enabling layer" — not a separate scam category, but a tool that makes every existing scam more convincing — is the detail worth passing along. A grandparent emergency call, an IRS threat, a romantic contact online: all of them are harder to spot when the voice, and increasingly the face, matches someone the victim already trusts.

Gobble's Take: The $16.6 billion figure is reported losses only — the actual number is almost certainly higher, since most victims never file a complaint.

Source: Vectr-Cast / cyberwarrior76

In Case You Missed It

Yesterday's top stories:

• Your Brain Fills in the Gaps: Why AI Voice Clones Don't Need to Sound Perfect
• Billions Lost, Organized Crime at Scale: The Infrastructure Behind the Fraud