🛡️ Staying Safe

~12 min read

Scams, privacy and knowing when AI is wrong.

**Spotting a Scam** — The tell-tale warning signs

Scam messages share tell-tale signs: a rush ('act now!'), a request for money or details, an odd link, or an offer too good to be true. You can even paste a suspicious message to the assistant and ask, 'does this look like a scam?'

• Watch for urgency, money requests, and 'too good' offers • Paste a suspicious message and ask if it looks like a scam • Real banks and services never rush you or demand secrecy

Did you know: A safe rule: no genuine bank, tax office or relative will ever pressure you to act 'right this second'.

**AI Can Be Wrong** — Trust, but verify

An assistant can state something false with complete confidence — this is called a 'hallucination'. It isn't lying; it simply guessed wrong. So for anything important — a fact, a date, a figure — treat its answer as a draft to check.

• AI can sound totally sure while being completely wrong • This mistake is called a 'hallucination' — it's common • Always double-check important facts, dates and figures

Did you know: The confident tone is the trap: a wrong answer sounds exactly as sure as a right one, so verify what matters.

**Checking a Fact** — A quick second look

When a fact matters, don't stop at the assistant's word. Ask 'how do you know that?' or 'can you search the web for this?', then confirm it with a trusted website. A minute of checking saves a world of trouble.

• Ask 'can you check that on the web?' for current facts • Confirm important claims against a trusted, known website • For official matters, always go to the official source

Did you know: Assistants that can search will show their sources — a good moment to click through and see for yourself.

**Guarding Your Privacy** — Keep private things private

Treat a chat like a postcard, not a locked diary — don't write anything you'd mind a stranger reading. Never type in passwords, full bank details, or your national insurance number. The assistant doesn't need them, and no honest task will.

• Never share passwords, PINs, or full bank card numbers • Skip your national insurance number and full address • If a task seems to need these, that's a red flag — stop

Did you know: A simple test before typing anything sensitive: 'would I write this on a postcard?' If not, don't type it in.

**What Never to Share** — The short forbidden list

Some details should never be typed into any chat, website or message, no matter who seems to be asking: passwords, PINs, full card numbers, and one-time security codes. Memorise this short list. Anyone demanding them is not to be trusted.

• Never share passwords, PINs or one-time security codes • No legitimate service asks for these in a chat or email • When in doubt, share nothing and ask someone you trust

Did you know: Banks sometimes deliberately test you — a real one hopes you'll refuse to read out your PIN, even to 'them'.

**Kinds of Online Trickery** — Name them, tame them

Trickery comes in familiar shapes: emails pretending to be your bank, calls claiming to be tech support, copycat websites, and strangers posing as friends. Naming them robs them of their power. When something feels off, it very likely is.

• Fake emails, calls, websites and 'friends' are the classics • Trust your instinct — 'this feels wrong' is worth heeding • Ask the assistant to explain any trick you've heard of

Did you know: The assistant is happy to describe the newest scams doing the rounds, so you know them before they knock.

**Pause Before You Trust** — The golden pause

The single best defence is a pause. When a message stirs fear or excitement, that's precisely the moment to stop, breathe, and check before acting. Scammers rely on speed; your calm is their undoing. There is always time to think.

• Any message urging haste deserves a deliberate pause first • Fear and excitement are the feelings scammers aim for • Ring the person or company back on a number you trust

Did you know: If a call claims to be your bank, hang up and ring the number on your card — never the one they give you.

**Real Photo or Fake?** — Seeing isn't believing

AI can now create convincing fake photos and even voices. A startling picture or a surprising video clip may not be real. Before you believe or share it, ask where it came from — a trusted news source, or a stranger online?

• AI can create fake photos, videos and even cloned voices • Look for odd hands, strange text, or too-perfect scenes • Check the source before believing or sharing anything

Did you know: If a shocking image or clip comes from a stranger rather than a known news source, treat it as unproven.

**Scams Are Getting Cleverer** — Stay a step ahead

Scams used to be full of spelling mistakes and easy to spot. With AI, they're now polished and personal, which is all the more reason to rely on habits, not appearances. A perfect-looking message can still be a trap.

• Modern scams are polished — no more obvious spelling errors • Judge by the request, not the neat presentation • Your safety habits matter more now than ever before

Did you know: A well-written, error-free message is no longer proof of honesty — if anything, it now warrants a touch more care.

**Your Safety Habits** — A calm little routine

Safety comes down to a few calm habits: pause when rushed, check what feels off, never share your secret details, and ask someone you trust when unsure. Keep these four close and you can explore with real confidence.

• Pause, check, protect your secrets, and ask when unsure • These four habits handle the vast majority of dangers • Being careful lets you enjoy AI freely, not fearfully

Did you know: There's no shame in double-checking with a friend or the bank — the confident, careful person is the safe one.