Issue No. 01

10 trail markers

Sunday, August 23, 2026 · Covering August 16–22

Voice-cloning scams, a $7.7 billion fraud toll, credit freezes, AI that books appointments, and how to tell a trending story from an important one.

Scams & money

That “emergency” call from a grandchild? It might be a fake voice.

Scammers can now use AI to clone a familiar voice — a grandchild, child, or spouse — and use it to ask for urgent money. It’s an active, escalating threat all through 2026.

If you get a call like this, the safest move is to hang up and call the person back on a number you already have saved, rather than trusting the voice on the line. A family passphrase set up ahead of time can settle a real emergency from a fake one.

Source: FTC Consumer Advice

Scams & money

Older adults reported $7.7 billion in fraud losses last year

Adults 60 and older filed more than 201,000 fraud complaints in 2025, with reported losses above $7.7 billion.

AI is making fake voices, emails, websites, and even video far more convincing than before. The core advice hasn’t changed: pause before acting, verify independently, and never pay with a gift card or cryptocurrency.

Source: National Council on Aging

Safety & risk

Five simple habits that block most AI scams

Treat any unexpected message from your bank, retirement fund, or a family member as unverified until you check it yourself.

Don’t click links in messages you weren’t expecting — go directly to the company’s official website or call the number on your card or statement. Worth remembering: credit cards generally offer better fraud protection than debit cards for online purchases.

Source: WTOP

Safety & risk

One free step that protects you from identity theft

Freezing your credit file with Equifax, Experian, and TransUnion is one of the strongest — and free — defenses against someone opening accounts in your name.

It’s worth doing even if you haven’t been targeted yet, and checking your credit report periodically helps catch problems early. Freezes can be temporarily lifted when you apply for something legitimate, then put back in place.

Source: CNBC

Rules & regulation

Washington is paying attention to senior-targeted fraud

The federal government has announced new efforts aimed at cross-border scam networks, specifically naming seniors as a group disproportionately harmed.

Good news, but worth remembering: policy takes time. The personal habits in the items above remain your best day-to-day protection while enforcement catches up to the scale of the problem.

Source: White House Fact Sheet

Everyday AI

AI can now help book your doctor and dinner reservations

Google’s Gemini assistant is adding connections to services like Zocdoc, OpenTable, and Angi/Thumbtack for home repairs.

For caregivers juggling a parent’s schedule, this could mean asking one assistant to help coordinate appointments instead of switching between apps. Before relying on it, confirm the booking went through directly with the provider.

Source: Google

Practical tools

A quick way to polish a letter or email with AI

If you use Claude for Word, highlight a paragraph and ask it to rewrite for clarity or shorten it, then review the suggestion before accepting.

Try being specific — for example, “rewrite this for my sister in plain language, under 100 words.” Always double-check names, dates, and numbers afterward, since AI can confidently get the small details wrong.

Source: Superhuman

Practical tools

An “always-on” AI assistant for repetitive tasks

A new tool called Grok Bot is designed to handle multi-step tasks — gathering information, drafting a summary, and sending it for review — without needing to be re-prompted each time.

For a caregiver tracking medications, appointments, or bills, this points to where AI assistants are heading. As with any new tool, set clear limits before giving it access to send messages or make purchases on your behalf.

Source: Superhuman

Language models

Claude will start “watermarking” AI-written text

Anthropic says future versions of Claude will add an invisible marker to text it generates, to help others tell that AI was involved in writing it.

It’s not foolproof — it can’t prove whether something was fully AI-written or just lightly edited — but it’s a small step toward more transparency as AI writing becomes more common, including in scam messages.

Source: Anthropic

Reading the news

How to tell if a “trending” AI story is really a big deal

When you see headlines like “AI is everywhere this week,” reporters count how many articles mention a topic over a set time period.

That count tells you how much attention something is getting — not whether it’s true, important, or good. Useful to keep in mind next time an AI headline feels overwhelming: attention is not the same as significance.

Source: Perigon