Latest additions — checked 29 September 2026
These September announcements add practical options for students and creators. The earlier brief below retains its original source-check date.
Connect creative and productivity apps to Gemini
Google announced a new Connected Apps rollout on 23 September, including Adobe, Picsart, Squarespace and Webflow for creative work, plus tools such as Airtable and Linear for projects. Check Gemini’s settings for the apps available to your account before planning a workflow. Official Google announcement.
MVR experiment: start with an outline for one college event poster. Check the words, dates and design yourself before sharing it.
Create across Google Workspace apps
Google’s 9 September announcement describes creating documents, spreadsheets and presentations across Workspace apps, using sources you select or your administrator enables. The rollout covers specified paid Workspace, Google AI and education plans; availability is not universal. Official features and eligible plans.
MVR experiment: turn a project brief into a draft presentation, then compare every slide with the original brief.
A September model update for builders
Google introduced Gemini 3.8 Flash on 2 September for coding, reasoning and multi-step agent tasks. Its announcement lists developer access through the Gemini API and Google AI Studio, and consumer access for Google AI Pro and Ultra subscribers. The separate Flash Cyber model has restricted access for trusted defenders. Read Google’s release and access details.
MVR experiment: ask for a small study-planner prototype using fictional data, then test its dates and calculations. Vendor performance claims are not a guarantee for your own task.
Source check: 27 September 2026. AI announcements are useful when they change something you can actually do. Here are a few recent developments, what they mean, and a small task to try. Availability can differ by account, plan and region.
1. Turn revision into active recall
OpenAI’s 22 September release notes announce interactive flashcards in ChatGPT. You can ask for a topic or provide notes, then mark answers you know and revisit the ones you miss. This is more useful when you test yourself before revealing the answer. Read the official release notes.
Try this: choose one chapter you already understand. Ask for ten questions using only your notes. Check the answers against your textbook, and rewrite any confusing cards in your own words. Keep a short list of concepts you still cannot explain aloud.
2. Voice is becoming a way to work
The 23 September ChatGPT update adds support for available plugins during Voice conversations on web and mobile. Access to Voice in Work depends on the account’s entitlements. Existing app permissions and limits still apply. Check the availability details.
Our suggestion: start with a low-stakes task, such as planning a college presentation. Speak your audience, time limit and main message. Review the resulting outline before sharing it. A fluent spoken response is not proof that every fact is correct.
3. Account security is easier to review
OpenAI’s 25 September notes introduce a security-history view for recent account activity. The web path is Settings → Security and login → Security history. Use the official help page if an event looks unfamiliar. See the announcement.
4. Practical AI learning is growing
Google’s AI Professional Certificate focuses on applying AI to workplace tasks such as research, communication and data analysis. It is a learning programme, not a promise of employment. Read Google’s programme announcement.
A better way to follow AI news
Before trying a new feature, ask: Is it available to me? Does it solve a real problem? What data would I be sharing? Can I check the result? Save one useful experiment rather than collecting ten subscriptions. This is a dated editorial brief, not a live news feed; revisit the linked sources for later changes.
Explore more ideas in Technology, or read how we approach our content.
