Plain-English guides for people meeting AI for the first time, written for readers who want the switch, the step or the setting – not the theory. Every price and menu path is dated, and we say plainly when the free option is the right answer.
On October 9, 2026 we read 9 Yahoo help pages: Yahoo labels 5 Mail features as AI and documents off steps for 3. The Message summaries switch on web and Android, what has no switch, and what Yahoo says about scanning mail.
On October 8, 2026 we read the home-phone help pages of 5 phone companies: none mention AI voices, but 4 screen robocalls by themselves. The FCC rule, *77 and *60 steps, and what changed with Nomorobo in 2026.
On October 7, 2026 we checked 67 YouTube videos whose titles say they were made with AI; only 24 carried the Made with AI label. Where the label shows, what Gemini can check, and five steps to try.
On October 6, 2026 we read 21 privacy pages from OpenAI, Google, Anthropic and Microsoft; 3 of the 4 makers say not to type sensitive information. What to leave out, the training switch, and Temporary Chat.
On October 5, 2026 we read 6 Amazon Kindle user guides; none describes a switch that turns off Ask this Book, Recaps, Story So Far or the Scribe AI tools. Five steps keep them out of your way.
On October 4, 2026 we read 8 Microsoft pages about Copilot in Edge; none describes one switch that turns off all AI. Step-by-step: turn off context clues, the Copilot new tab page, Journeys, the Copilot button, writing help and Copilot memory.
On October 3, 2026 we read Meta’s Facebook Help Center and counted 9 AI features; only 2 have a real off switch. Step-by-step: mute the Meta AI chat, turn off AI images of you, remove comment summaries, switch dubbed reels to original audio, and object to AI training.
On October 2, 2026 we read Microsoft’s support page and the Google page-one results: the Enable Copilot checkbox exists only in Word for Windows and Mac, and each app and computer needs its own. Step-by-step for both, what to do if the box is missing, and why not to roll back Word.
On October 1, 2026 we read nine official U.S. pages on family emergency scams: 4 recommend a family safe word, and all 9 say to verify first through a contact you start. How to pick a word a cloned voice cannot guess, and why not to copy any example online.
On September 30, 2026 we read 13 help pages from Zoom, Otter and Read AI and counted 8 kinds of AI that can listen to a Zoom call. Any guest can remove Otter and Read AI with a chat command; Zoom’s own summary is switched off only by the host.