The most important news pieces from the AI world are here for you, whether you’re an AI enthusiast or just wish to keep an eye on what’s going on.

Focus Topic: Microsoft pushes for relevance with its “human-centered AI” release

Microsoft held its annual Fall event last week, rolling out a wave of updates to its Copilot platform. The focus? A new narrative around “human-centered AI,” positioning Copilot not just as a tool, but as a digital companion designed to empower people — not replace them. The company is clearly aiming to stay front-of-mind in the fast-moving AI race.

While the messaging emphasizes trust, support, and usefulness, many of the announced features echo what other players already offer. Rather than a technological leap, this release feels more like Microsoft trying to catch up and reframe the conversation. Still, there are meaningful upgrades worth knowing about.

Here’s a quick overview of what’s new:

Mico avatar – a customizable visual companion that reacts with tone, color, and animation during conversations
Memory & personalization – lets Copilot remember key info across sessions (like tasks or preferences), with user control over editing and deletion
Proactive Actions – enables Copilot to suggest next steps or insights based on user activity or past queries
Connectors – integrate data from services like OneDrive, Google Drive, Outlook, Gmail to make information searchable across tools
Copilot Groups – shared AI chats for teams (up to 32 people) with summarization, task-splitting, and collaboration features
Health & education tools – improved health answers from verified sources (like Harvard Health), and Socratic-style tutoring with voice + visuals
Edge Copilot Mode – browser integration with voice navigation, auto-summarization, and “Journeys” to organize past browsing

Copilot in Windows – system-wide AI integration with wake word access, Copilot Vision for file guidance, and app recall
Copilot Pages – multi-file workspace for AI-driven collaboration
Copilot Search – combines web links with AI-generated answers in a single view

In short, Microsoft is expanding Copilot from a chat assistant into an all-around AI layer across its ecosystem — with a stronger emphasis on design, memory, and integration. Whether it actually becomes your “everyday AI companion,” as Microsoft suggests, remains to be seen.

LLMs & AI Models

- OpenAI acquired Software Applications Inc., the team behind Sky – a Mac overlay tool that automates tasks by reading and acting on screen content.
- ChatGPT added company knowledge support, letting it reference internal docs from Slack, Google Drive, GitHub and cite them directly in responses.
- OpenAI launched Atlas, an AI-powered browser for Mac that recalls past activity, navigates autonomously, and supports Agent mode.

- Atlas also raised concerns after researchers found a prompt injection vulnerability that allowed sites to manipulate user clipboard content.
- Claude’s memory is now live for Max users and coming to Pro soon, offering project-based context handling for more personalized outputs.
- Gemini AI now has real-time map awareness, pulling live data like opening hours and ratings for 250M locations via Google Maps.

New Tools

- Lightricks released LTX-2, a 4K AI video model that generates up to 15-second synced video clips with natural resolution and audio.
- Tencent open-sourced Hunyuan World 1.1, a 3D generator that recreates objects from photos or video using only one GPU.
- Nucleus Genomics launched Origin AI, an IVF-focused model predicting embryo disease risks by analyzing 7M genetic markers.
- Runway added fine-tuning tools for its video models, allowing users to customize generative video content with their own data.
- Manus upgraded its agent platform to version 1.5, delivering 4x faster execution and full web development capabilities.
- Google launched Google Skills, an AI learning hub with 3000+ free courses, labs, and certificates for teams and individuals.

- Lovable introduced a Shopify integration that lets users build online stores via natural language prompts in multiple languages.
- Claude Code is now available in-browser, supporting cloud-based coding, real-time progress tracking, and GitHub integration.
- Napster launched Napster 26, a $99 holographic AI platform that projects 3D virtual assistants above Mac screens without glasses.
- YouTube unveiled a deepfake detection tool for creators to report and remove AI-generated copies of their faces or voices.

Other Quick Picks

- Netflix confirmed it’s going “all in” on AI, applying it in content production, visual effects, and internal processes.
- Amazon introduced AI-powered smart glasses for couriers, showing navigation, package info, and safety alerts in real-time

- Meta cut around 600 jobs in its AI division, affecting research and infrastructure teams, but kept its superintelligence lab intact.
- Anthropic responded to US government criticism by reaffirming Claude’s neutrality and compliance, citing federal use and contracts.
- Channel 4 aired the UK’s first AI TV host, only revealing it was artificial at the end of the show.
- OpenAI tightened Sora rules after backlash from Hollywood over unauthorized celebrity likenesses used in AI videos.
- Andrej Karpathy downplayed the current hype around AI agents, saying general-purpose autonomy is still at least a decade away. Watch HERE.
- AI eye implants helped patients with vision loss to read again, using infrared signals from smart glasses to stimulate the retina.
- The Future of Life Institute published an open letter urging governments to pause the development of artificial superintelligence (ASI) until safety is proven.

🇪🇪 AI News from Estonia

- Kristiina Polli, head of Tagasi Kooli, warned that AI could become an unquestioned authority for students, urging focus on critical thinking over facts.
- Dutch artist Jeroen van der Most opened an AI-powered art exhibit in Tallinn, using “vibe coding” to turn Excel sheets into generative floral artworks. Read here.

- Silver Keskküla‘s startup ScamGuardian.ai partnered with Tele2 Estonia to block scam calls using AI trained on 12,000 daily call attempts.
- Andrus Ait, CFO at Port of Tallinn, shared how he uses ChatGPT and Claude for analysis, but emphasized caution and rule-based AI use.
- Riigi Infosüsteemi Amet // Estonian Information System Authority (NCSC-EE) called on leadership to take direct responsibility for security awareness and not leave it to IT teams alone. Read more here and check out also itvaatlik.ee.
- Lauri Luht, Estonia’s new national CIO, emphasized the country’s potential to become a global testbed for AI-driven digital government services. Read here.
- 🎙️ AIPowerment Podcast released a new episode featuring Mariana Hagström and Vladislav Šikirjavõi from Avokaado, discussing AI in legal document workflows.

🎙️ AIPowerment Podcast dropped a new end-of-month episode featuring Mariana Hagström and Vladislav Šikirjavõi from Avokaado, discussing AI in legal document workflows.

🎧 Listen to AIPowerment Podcast on Spotify, Apple Podcasts, and YouTube.

Want to stay in the loop?

Straight to your inbox: practical AI updates, finance use cases, tools to try, and upcoming trainings.
I cut the noise and send only what’s actually useful.

💼 Exploring how to make AI work in finance? Let’s talk practical use cases – connect with Gerlyn Tiigemäe for expert guidance.

📚 If you would like to participate in one of my trainings or listen to speaking engagements, here are the upcoming ones:

Training “AI võimalused finantsvaldkonnas” @ Äripäeva Akadeemia, last free spots for December – information
Training “AI võimalused projektijuhtimises” @ Äripäeva Akadeemia, next trainings in 2026 – information
Training “AI uued rakendused”, 12.11.2025 – information
Training “AI assistendi roll ja võimalused”, 26.11.2025 – information

Your Weekly AI news roundup @27.10.2025