Google Takes On Canva with “Prompt-to-Design” Tool

Google is making a major push into the creative software market with Google Pics, a new AI-first design tool for Workspace users. Instead of traditional design interfaces, users describe what they want, and the tool—powered by Gemini and the “Nano Banana” model—generates and edits “professional-grade” images. Industry observers see it as a direct challenge to Canva and Adobe’s dominance.

• Google’s answer to Canva is an AI tool where you prompt instead of design (TechCrunch)
• Try Google Pics: Easy image creation and editing in Google Workspace (Google Blog)
• Google Pics is like Canva, but with even more AI (The Verge)

Separately, Google DeepMind announced “agentic video understanding” for Gemini, allowing the model to take multi-step actions within videos, like extracting specific information or summarizing content based on complex queries.

• Introducing agentic video understanding with Gemini (Google DeepMind)

Healthcare Gets an AI Boost

OpenAI’s ChatGPT Health is now integrated with Epic, the largest electronic health record system in the U.S. The integration gives clinicians read-only access to patient data within ChatGPT, aiming to streamline workflows. OpenAI also published a broader announcement about connecting ChatGPT to trusted healthcare data sources.

• ChatGPT Health adds Epic integration for clinicians to import patient data (TechCrunch)
• Healthcare organizations can now connect EHR and additional industry data to ChatGPT (OpenAI)

The Rise of the Specialized Agent

A clear trend emerged today: highly specialized AI agents built for specific professions and use cases.

• For Families: Fambot launched an “AI chief of staff” to manage family logistics like calendars, school emails, and sports schedules.
• For Farmers: John Deere is testing “JD,” a chatbot that uses a farmer’s own field and machine data to answer questions about equipment settings, harvest timing, and best practices.
• For Police: Startup Blue Voice raised $6M to build a “Harvey for police officers,” an AI assistant trained on department-specific laws and protocols not available on the public web.
• For Infrastructure: Empirik, incubated by Sequoia, launched with $21M to predict IT outages before they happen, aiming to be the “Cursor for infrastructure.”
• For Enterprises: OpenAI highlighted how “AI-native” companies like Basis and Clay are using AI agents to transform internal workflows for onboarding and account management.

Policy, Security & Funding

• National Security: The Pentagon has added custom, secure versions of ChatGPT and Grok to its central AI tools portal, joining Google’s Gemini.
• Platform Policy: Instagram is placing new limits on the reach of undisclosed AI-generated influencer profiles in response to user frustration.
• Corporate Espionage: Apple presented what it calls “shocking evidence” against a former employee accused of stealing company data for OpenAI, alleging he destroyed evidence.
• Agent Security: Startup AIR raised $50M for a platform that helps companies discover, vet, and monitor the skills used by AI agents within their organizations to block unwanted behavior.
• Media Search: AI video search startup Clipto hit a $250M valuation after reaching profitability, showing strong demand for tools that can search through terabytes of video content.

In Other News

• Nvidia is set to launch its controversial DLSS 5 AI upscaling tech this week, which will require powerful new RTX 50-series GPUs.
• Amazon Alexa is testing a new “Update Me When” feature that sends personalized alerts about product launches or events likely to tempt users to shop.
• MIT PhD student Ila Kumar is researching how to give young people an active role in shaping the digital technologies that affect their well-being.
• A warning from Ars Technica advises caution with media streaming devices that offer free content, as they can secretly turn home networks into proxy nodes for potentially malicious traffic.

Editorial: The Age of the Agent

Today’s news underscores a pivotal shift: AI is moving beyond general-purpose chatbots into a proliferation of specialized agents. Whether it’s a “chief of staff” for a busy household, a copilot for a farmer, or a secure assistant for a soldier, the value is increasingly found in deep, context-aware integration with specific data and workflows. This specialization brings immense potential for productivity but also new layers of complexity around security, ethics, and transparency—as seen in the funding for agent-vetting startup AIR and Instagram’s new rules for AI profiles. The race is on not just to build the most powerful model, but to build the most useful and trustworthy agent for the job.