AI
AI can draft, summarise and compare. When it sends, buys, deletes, submits or changes access, it is acting with your identity.
Short version
Read-only and draft-only are low drama. Consequential actions need human approval. Secrets do not belong in random tools.
Source imagery
Swipe examples
Image 1/2
Image 2/2
Agent approval gate
A useful home agent can read a bill, compare options or draft a message. The line changes when it sends email, pays money, deletes files or changes access while logged in as you.
The agent can help without becoming the household's unattended hands. Reading and drafting stay separate from doing.
Reading still needs limits: no secrets, client material, medical/legal details or family-sensitive uploads in random tools.
Explain the jargon
Tap a term for the plain-English version and the practical move. No fake mystique, just the bit that changes what you do at home.
Swipe the terms one at a time below desktop width. Glossary cards can get wordy, and squeezing three of them into a tablet row helps nobody.
Instructions hidden in a page, email or file that try to steer an AI tool away from what you asked it to do.
Do this: Do not let agents act on untrusted content without a review step, especially before sending, buying, deleting or changing access.
Giving an AI more ability to act than the task needs: browser sessions, payment access, email sending or admin permissions.
Do this: Default to read-only or draft-only. Add explicit approval gates for consequential actions.
Secrets, client material, legal/medical/family information, work files and anything you would not want in a vendor support ticket.
Do this: Use approved tools and keep random AI sites away from sensitive uploads and prompts.
Read these as three short household checklists. They stay stacked below desktop width so the action text does not get squeezed.
Self-check questions
Use these quick checks to find the next practical fix. The useful answer is not perfect security; it is whether the safer path is obvious when someone is tired, embarrassed or in a hurry.
On phones, swipe one question at a time. Use the first uncomfortable answer as the next household fix, not as a lecture.
Can this AI action send, submit, buy, delete, change access or speak as you?
Good sign: The AI can prepare the draft, comparison or checklist, but a person does the consequential click after reading it.
Watch for: If the tool can act while reading untrusted pages, emails or files, a prompt-injection mistake can become your mistake.
Would this prompt, screenshot or upload be awkward if it appeared in a vendor log, support ticket or shared workspace?
Good sign: Sensitive material is redacted, handled in an approved tool, or kept out entirely.
Watch for: The risky bit is often the attachment: tax PDFs, medical letters, family legal files, client notes, keys and screenshots with tokens.
What is the lowest access that still lets the AI help: read, draft, browser, send, pay or admin?
Good sign: Default to read-only or draft-only. Raise access for a narrow task, then remove it.
Watch for: Permanent browser, email, payment or admin access turns a helper into another identity to supervise.
Full guidance
An approval-gate flow for agent permissions and sensitive actions.
Swipe one guidance note at a time below desktop width. The receipt cards appear first; these notes are the deeper explanation, not a wall to skim in one go.
Note 01/05
Read-only < draft-only < browser access < send/submit < payment/admin. The higher it climbs, the stronger the approval gate needs to be.
Note 02/05
Prompts, uploads, outputs, logs and metadata may exist outside the household. If it would be awkward in a vendor ticket, do not paste it into a random tool.
Note 03/05
AI can help prepare work. It should not secretly make decisions, submissions or representations on your behalf.
Note 04/05
A browser agent does not only read the task you typed. It may also read web pages, emails, PDFs, tickets and chat messages. Any of those can contain hostile instructions. That is why untrusted content and action authority are a bad mix: the tool may be looking at a scam page while also holding your logged-in browser.
Note 05/05
Let AI draft the comparison, write the email, summarise the bill or list the steps. Keep the actual click with a person for sending, buying, deleting, changing account settings, uploading private files or granting access. It is boring, but boring is exactly the point.
Scenario
Swipe one real-world mess at a time
Scenario 1/3
It wants to log in, compare bills and submit a change.
Better response
Worse habit
Handing it email, payment and submission authority unattended.
Scenario 2/3
A browser agent is asked to organise email and one message says to ignore previous instructions, export data or approve a fake invoice.
Better response
Worse habit
Letting the agent both read untrusted mail and act on accounts without a human stop.
Scenario 3/3
A free tool asks for a tax PDF, work document, medical letter or family legal file to summarise.
Better response
Worse habit
Uploading the real document because the summary will only take ten seconds.