New AI models, app changes and new courses. Short and to the point, so you don't miss anything important.
App changeChatGPT22 October 2026
You can now hide selected ChatGPT conversations from your history, pinned chats and search, and open them only with a 4-digit PIN or your phone's unlock. It's rolling out gradually, including to the free plan. Handy when you share your screen or lend your computer, but it isn't a vault.
- lock a chat via ••• → Lock chat; find them under Projects → Locked chats
- locking doesn't change how long OpenAI keeps the chat or your training setting
- if you forget and reset your PIN, locked chats are deleted for good
- don't put sensitive data into AI, not even into a locked chat
Privacy and AI courseApp changeClaude22 October 2026
Anthropic has taken the beta label off Docs, Slides and Design in Claude. You can ask for documents, decks and designs in a chat on every plan, including free. There are also two new beta features: live dashboards from company data on paid plans and Motion animations for teams.
- download a deck to PowerPoint or PDF and it looks like it did in the editor
- decks also go straight to Google Slides as editable files
- you can edit a doc, deck or design in the Claude mobile app
- if you use the standalone claude.ai/design site, move your work by December 14, when it closes
Claude in depth courseNew modelChatGPT18 October 2026
OpenAI is rolling out a new model, GPT-Live-1.1, in voice conversations, with more natural accents across languages. The free version gets the smaller GPT-Live-1.1 Quip with limited time, Plus gets 3 hours per 24 hours with the full model.
- pick a voice in Settings → Voice → Voice
- new voices include German and French options, among others
- you can still talk in your own language, or ask it to switch
- camera and screen sharing only work with the Advanced option in the mobile app
ChatGPT in depth courseApp changeClaude18 October 2026
Anthropic has launched the Claude for Google Workspace add-on. Claude opens in a sidebar next to the file you have open and edits it right there, with no copying back and forth. It's a beta for the paid Pro, Max, Team and Enterprise plans.
- install it once from the Google Workspace Marketplace and it works in Docs, Sheets and Slides
- open it in a file via Extensions → Claude → Open Claude
- it only sees the open file, not your whole Drive or email
- every change it makes shows up in version history, so you can undo it
Claude in depth courseNew modelGemini14 October 2026
Since October 13, Google has been adding a new model, Gemini 4 Argon, to the Gemini app. It's built for complex tasks from start to finish. Google AI Pro and Ultra subscribers get it worldwide.
- Google says it answers more directly, without needless repetition or clichés
- on longer tasks it keeps you posted on its progress
- for now Google offers it only on the AI Pro and Ultra plans
- it's rolling out gradually, so it may take a few days to show up in your model menu
Gemini in depth courseNew courseChatGPT12 October 2026
The third level after ChatGPT in depth. In ten days you turn ChatGPT into a system that works for you: projects that know your work, recipes for tasks you do over and over, and tasks that run without you.
- memory and projects you don't have to brief every time
- plugins instead of custom GPTs, which end in December
- permissions: what ChatGPT may do on its own and what never
- scheduled tasks that watch for changes and know when to stop
- most of it works for free; Work and email-triggered tasks need Plus
Advanced ChatGPT courseApp changeGemini12 October 2026
Google has clarified how Gemini works without a subscription. You don't pick the model yourself, Gemini uses Auto. Most questions go to the fast Flash-Lite model, and harder ones may go to the stronger Flash or Pro.
- without a subscription, Auto is the default
- if you turn off automatic model selection, Flash-Lite answers everything
- Google AI Plus: Flash-Lite and Flash, AI Pro and Ultra: all three models
- harder tasks use more of your limit, which refreshes every 5 hours
Gemini in depth courseApp changeGemini9 October 2026
Google has opened SynthID Detector to everyone. Upload an image, video or audio file and find out whether it was made with AI from Google or partners like OpenAI. For now the tool is in English only.
- it spots the hidden watermark from Google, OpenAI, NVIDIA and Kakao AI, and later Apple
- you can also check right in the Gemini app, Google Search and Chrome
- no watermark found doesn't mean it's real: AI from other companies may not add one
- for a suspicious video, still ask: who posted it, and do trusted news outlets report it?
Safe with AI courseNew course9 October 2026
Two short guides are here, and you don't even need to sign in. The Install AI guide walks you step by step from choosing an AI to your first sign-in on computer and phone. Secure your account has 3 parts of 10 minutes each.
- Install AI: pick your system and AI, download it from the official source and sign in
- Secure your account: two-step verification, signing in without a password (passkeys) and checking for unknown devices
- find them in the menu under Tutorials
Install AINew modelChatGPT8 October 2026
OpenAI is bringing GPT-6 to everyday chat. Paid plans get it from 7 October, the free version and Go gradually from 8 October. The new part is Intelligent UI: ChatGPT can answer with pictures, side-by-side comparisons, charts or a small tool like a calculator, not just text.
- free and Go: GPT-6 Luna, paid plans: GPT-6 Sol
- ask for a calculator or a bill splitter right in the chat
- it starts answering sooner, while it's still thinking
- double-check numbers that matter
ChatGPT in depth courseApp changeChatGPT7 October 2026
You can now attach an audio recording to ChatGPT, like a meeting, an interview or a lecture. You'll get a transcript, a summary or a draft email, and you can ask what was said. It works on paid plans, not on the free version yet.
- MP3, M4A, WAV and more, files up to 512 MB
- video files don't work, audio only
- transcripts can contain mistakes, check important details in the recording
- only upload recordings the people in them have agreed to
ChatGPT in depth courseApp changeGemini6 October 2026
Google is changing which models you can use in Gemini. Without a subscription, from Thursday 9 October Auto mode picks the model for you, mostly the fast Flash-Lite. You can only choose the stronger models yourself on paid plans.
- no subscription: Auto mode (mostly Flash-Lite)
- Google AI Plus: Flash-Lite and Flash, no more Pro
- AI Pro and Ultra: all models, as before
- applies to personal Google accounts
Gemini in depth courseApp changeChatGPT6 October 2026
Over the coming weeks, OpenAI will start adding an invisible watermark to ChatGPT text in the European Union. The EU AI Act requires it. You can't see it, only a special tool can detect it.
- all plans, but only in the EU
- the watermark doesn't reveal who wrote the text or your chat
- editing the text weakens the watermark
- the detection tool isn't public yet
Safe with AI courseNew courseGemini5 October 2026
The third level after Gemini in depth. In ten days you turn Gemini into a system that works for you: skills for tasks you do over and over, your own sources and scheduled actions that run without you.
- from Gems to skills, and how to write a good skill
- connected apps and what not to hand over to them
- scheduled actions and Canvas for family or colleagues
- you need a personal Google account, no subscription required
Advanced Gemini courseApp changeCopilot5 October 2026
When you ask Copilot in Word a question, the answer now shows its sources: web pages and your work files. One click and you can check whether it's right.
- for Word in work Microsoft 365
- click the source for any number that matters
- Copilot in Word can now also arrange and edit images and shapes
AI in Microsoft 365 courseApp changeGrok5 October 2026
Grok now has skills. Describe once how you want something done, such as a shopping list sorted by supermarket aisle, and Grok uses it by itself next time. It works on the free plan too. We've added it to day 3 of the Grok in depth course.
- find them in the text box: “+” → Skills
- create a skill in a chat, write it manually or upload a file
- every account has built-in skills for Word, presentations, spreadsheets and PDFs
- when Grok uses one, you'll see “Loaded skill” in the chat
Grok in depth courseApp changeChatGPT4 October 2026
Clothes and accessories in ChatGPT now have a “Try on” button. Send a selfie and ChatGPT shows how the item could look on you. On iPhone you can photograph several pages in a row and ChatGPT turns them into one PDF.
- scanning: camera in ChatGPT → three dots → Scan (iPhone for now)
- ChatGPT saves your selfie for future try-ons
- change or delete it in Settings → Personalization → Reference photos
- try-on may not be available in every country yet
Privacy and AI courseApp changeGemini1 October 2026
Google has launched skills: saved instructions for tasks you do over and over. On personal accounts they replace Gems from November. You won't lose your Gems, Google converts them into skills for you.
- call up a skill with a slash (/) in any chat
- you can add files to it, such as a PDF or an image
- for now only for people aged 18 and over
- work accounts keep Gems until March 2027 at the earliest
Gemini in depth courseNew courseChatGPT30 September 2026
The second level after ChatGPT from scratch. In ten days you learn to let ChatGPT work on bigger tasks, with your files and on a schedule without you. Most of it works with the free version.
- projects, files and spreadsheets, deep research
- plugins for Gmail and Calendar, and safe permissions
- scheduled tasks and Work mode
- first day free
ChatGPT in depth courseApp changeChatGPT30 September 2026
Much like “Sign in with Google”, some services now offer sign-in with your ChatGPT account. One less password to remember, but your ChatGPT account then becomes the key to other services too.
- the service gets your name, email and profile photo
- your chats and memory are not shared
- any further access needs your separate approval, so read carefully what you confirm
- turn on two-factor authentication in ChatGPT
Safe with AI courseNew modelChatGPT30 September 2026
OpenAI announced more than 20 new features. Most of them are in the pricier Pro plan or in business plans. If you're on Free, Go or Plus, almost nothing changes in everyday chat for now.
- new model GPT-6.1 Sol in Work mode, gradually coming to Plus
- dots agents work for you around the clock, but not yet on the Pro plan in the EU
- Space and Pages for working together are in Pro and business plans
- new Pro plan at 500 USD a month
ChatGPT from scratch courseNew modelClaude29 September 2026
Anthropic has added Sonnet 5.5 alongside Opus 5.5. It's faster and cheaper, and it's enough for everyday tasks: emails, documents, presentations and spreadsheets.
- pick the model by clicking its name next to the send button
- good for everyday tasks and creating documents
- keep Opus 5.5 for the harder things
- Haiku 5.5 came out on 7 October, mainly for developers
Claude in depth courseApp changeGemini28 September 2026
From 29 September, some Google business plans turn on meeting notes automatically when at least three people join. If you don't want that, change it in the Meet settings or ask your IT admin.
- on by default in Business Standard and Business Plus
- off by default in Enterprise plans
- notes work in English and 7 more languages
- personal Google accounts are not affected
AI in Google Workspace courseApp changeChatGPT26 September 2026
ChatGPT has added a security history. It shows sign-ins and security changes to your account, with time, place and device. If you see anything you don't recognise, change your password straight away.
- on the web: Settings → Security and login → Security history
- you see sign-ins, sign-outs and changes to two-factor authentication
- for the free plan and paid plans
Safe with AI courseApp changeChatGPT25 September 2026
ChatGPT is gradually rolling out the Privacy Center: an overview of what happens to your chats, memory and data, with links straight to the settings.
- on the web: account menu → Help → Privacy center
- on mobile: Settings → Privacy Center
- it changes nothing by itself, it just shows where to set what
- for the free plan and paid plans, rolling out gradually
Privacy and AI courseNew modelClaude25 September 2026
Anthropic released Claude Opus 5.5 on 22 September. It's faster and better at writing, working with documents and longer tasks.
- Anthropic also raised the limits on the Pro, Max and Team plans
- it writes answers more than a third faster than its predecessor
- its performance comes close to the strongest Claude models
Claude from scratch courseNew modelGemini25 September 2026
Images in Gemini are now made by the Nano Banana 2 model. Just write “Create an image…” or open Images in the side panel.
- more accurate text in images, in other languages too
- edit just a selected part of an image
- the same character across several images
- free at 1K resolution, 2K with a Google AI plan
Gemini from scratch course (images are on day 7)App changeCopilot24 September 2026
Since 18 August 2026 there is one Copilot app for the web, computer and phone, for personal and work accounts.
- sign in with a Microsoft, Apple or Google account
- according to Microsoft, your chats and files are not used to train AI
- podcasts, Deep Research and group chat have ended
- files from the app are saved to OneDrive
Copilot from scratch courseNew course24 September 2026
Three new ten-day courses for complete beginners. The first day of each is free.
- ChatGPT from scratch: the most widely used AI, at home and at work
- Copilot from scratch: Microsoft's AI, close to Word and Excel
- Gemini from scratch: for people who live in Google
See the courses