Better fakes.
The same right to pause.

What is already changing

Generative tools can produce convincing text, images and audio at low cost. The NCSC describes their use in impersonation and more convincing phishing. That does not mean every polished email is a scam, or that an awkwardly written one is harmless.

A personalised detail about your job or family is not proof of identity. It may be public, copied from social media or obtained from leaked information.

What we expect next

Our assessment, not a certainty: more tailored messages and easier voice imitation will make independent verification more valuable. Screening and identity checks may improve too, but availability and effectiveness will vary. We are not assuming any future technology will make spam disappear.

Can AI check your messages?

It may help summarise a long message or point out a suspicious request. It can also make mistakes, overlook a scam or invent reassurance. A “percentage written by AI” is not a useful safety verdict. Human-written messages can be fraudulent; AI-assisted writing can be perfectly legitimate.

Avoid pasting passwords, account numbers or other people’s private information into a tool. Before connecting an inbox, review what it can read, send, retain or share. Use work-approved tools for work email. Keep human review for money, identity checks and account changes.