Search

AI News January 28, 2026: Gemini in Chrome, OpenAI Prism, Claude interactive tools

AI News January 28, 2026: Gemini in Chrome, OpenAI Prism, Claude interactive tools

Major Week for AI Interfaces

From January 26 to 28, 2026, the three AI giants (Google, OpenAI, Anthropic) are launching products that integrate AI directly into everyday tools: browser, office suite, and professional apps.


Google: Gemini Integrated Everywhere

Gemini in Chrome with Auto Browse

January 28 โ€” Google launches a major redesign of Gemini in Chrome. The AI assistant now appears in a permanent side panel and can automatically browse the web.

New Features:

FeatureDescription
Side panelGemini permanently accessible next to the active tab, without switching tabs
Nano BananaImage transformation directly in the browser (room redesign, infographics)
Connected AppsIntegrations with Gmail, Calendar, YouTube, Maps, Shopping, Flights
Auto browseMulti-step automatic navigation: hotel search, form filling, price comparison
UCPUniversal Commerce Protocol, open standard co-developed with Shopify, Etsy, Wayfair, Target

Auto browse is the flagship feature: the user gives a complex task (book a flight, find an apartment, collect tax documents) and Chrome executes it by navigating across multiple sites. The agent requests confirmation before sensitive actions (purchases, sending messages).

Available for AI Pro and Ultra subscribers in the United States.

๐Ÿ”— Gemini in Chrome


Agentic Vision in Gemini 3 Flash

January 27 โ€” Google adds Agentic Vision to the Gemini 3 Flash API. This capability transforms image understanding into an agentic process: the model can combine visual reasoning and code execution to analyze images iteratively.

First supported tool: Python code execution to process visual analysis results.

๐Ÿ”— Agentic Vision @GoogleAI


Google AI Plus Available in the USA

January 27 โ€” Google launches Google AI Plus in 35 new countries including the United States for $7.99/month.

IncludedDetail
Gemini 3 ProIn the Gemini app
Nano Banana ProImage generation
FlowAI video creation tools
NotebookLMResearch and writing
Storage200 GB
Family SharingUp to 5 members

Existing Google One Premium 2TB subscribers automatically receive AI Plus benefits. Launch promo: 50% off for the first 2 months.

๐Ÿ”— Google AI Plus


OpenAI: Prism for Scientists

Prism: Free Workspace Powered by GPT-5.2

January 27 โ€” OpenAI launches Prism, a free workspace for collaborative scientific writing, directly integrated with GPT-5.2 (their best model for mathematical and scientific reasoning).

What Prism Does:

FeatureDescription
Native LaTeX WritingComplete cloud environment, no local installation required
GPT-5.2 in ContextThe model has access to the entire article (equations, citations, figures, structure)
Unlimited CollaboratorsNo seat limit, real-time work
Integrated SearchSearch on arXiv and direct integration into the manuscript
Whiteboard to LaTeXConversion of hand-drawn equations into LaTeX
Voice EditingSmall edits without interrupting writing

Origin: Prism is based on Crixet, a cloud LaTeX platform acquired by OpenAI.

Availability: Free for all personal ChatGPT accounts. Advanced features coming for Business, Team, Enterprise, and Education plans.

๐Ÿ”— Prism | Announcement


Anthropic: Interactive Tools and UK Partnership

Interactive Tools in Claude

January 26 โ€” Claude can now display and interact with tools directly in the conversation. No need to switch tabs to use professional apps.

Available Interactive Apps:

AppPossible Actions
AsanaCreate projects, tasks, timelines directly
SlackSearch conversations, preview and send formatted messages
FigmaTransform text/images into FigJam diagrams
AmplitudeBuild and explore analytics charts
BoxSearch files, preview, extract insights
CanvaCreate presentations with custom branding
ClaySearch companies/contacts, write outreach
HexData questions with interactive charts and tables
monday.comManage projects, assign tasks, visualize progress

Technology: Based on MCP Apps, a new extension of the Model Context Protocol that allows any MCP server to provide an interactive interface.

Available on web and desktop for Pro, Max, Team, Enterprise plans. Coming soon to Claude Cowork.

๐Ÿ”— Interactive tools in Claude | MCP Apps directory


Claude Free Plan: Files, Skills, Compaction

January 26 โ€” Anthropic extends features to free users:

FeatureDescription
FilesCreate and edit files (Excel, documents, PDF)
SkillsAccess to Claude skills
CompactionClaude continues working on long tasks

๐Ÿ”— @claudeai on X


UK Government Partnership: Claude on GOV.UK

January 27 โ€” Anthropic signs a partnership with the UK Department for Science, Innovation and Technology (DSIT) to build an AI assistant on GOV.UK.

Initial Use Case: Job support โ€” personalized advice for finding work, accessing training, understanding available support.

Key Features:

  • Agentic system that actively guides (not just answering questions)
  • Persistent context between sessions
  • Full user control over data (opt-out possible)
  • Anthropic engineers work with civil servants for skills transfer

This partnership follows the Memorandum of Understanding signed in February 2025. The goal: for the UK government to maintain the system independently.

๐Ÿ”— Anthropic Announcement


What This Means

This week marks a turning point: AI is leaving dedicated chat interfaces to integrate directly into everyday tools.

  • Chrome becomes an agent capable of executing multi-site tasks
  • Claude displays professional apps within the conversation
  • Prism integrates GPT-5.2 directly into the LaTeX editor

The common message: AI should be where you work, not in a separate tab.

On the business model side, OpenAI makes a bold bet by making Prism free โ€” aiming to become the default tool for the scientific community before monetizing advanced features.


Sources