OpenAI is creating an AI-powered hiring platform launching mid-2026 to connect businesses and workers, competing with LinkedIn. It will include a track for small businesses and local governments. OpenAI will offer AI fluency certifications aiming to certify 10 million Americans by 2030, supporting White House AI literacy efforts amid concerns about AI disrupting jobs.
Create professional Google Ads performance reports in Google Slides with AI insights and visualizations. Transform Google Ads data into polished presentations featuring AI-generated insights and custom visuals to impress stakeholders and streamline reporting. Use cases include reducing manual reporting time, presenting insights without design skills, maintaining reporting schedules, identifying trends, and sharing campaign insights for strategic decisions.
GeminiApp now supports uploading any file type, including highly requested audio files, within the existing multi-file workflow for documents and images. Users can attach up to 10 files per prompt, and ZIP archives are supported, making it easier to upload raw tracks or multiple interview takes together.
AI Mode is now available in Hindi, Indonesian, Japanese, Korean, and Brazilian Portuguese. This expansion uses advanced Gemini 2.5 technology to provide nuanced, locally relevant AI search experiences beyond simple translation. It allows users worldwide to ask complex questions and explore the web deeply in their preferred language. AI Mode can be tried at google.com/ai.
Apple plans a 2026 update for Siri and Apple Search with an AI web search tool called World Knowledge Answers, summarizing internet info for faster, accurate results. It will be added to Safari, Spotlight, and iPhone search. The tech, Linwood and LLM Siri, is part of the Luck E update, possibly iOS 26.4. Apple is working with Google, using a Google AI model for voice assistant and search, including text, photos, video, and local points of interest.
Google Vids is adding new generative AI features, including turning images into short videos using Veo 3. These features will be available to paid Google Workspace customers and Google AI Pro and Ultra subscribers, along with AI avatars and automatic transcript trimming. The basic Vids editor, without AI, will now be free for consumers. A new instructional series called "Vids on Vids" is also available to help users learn the tool.
Anthropic is testing Claude for Chrome, an AI that helps with tasks in the browser like managing calendars and emails. The pilot involves 1,000 trusted users to address safety risks, especially prompt injection attacks where malicious instructions trick the AI. New defenses cut attack success rates but more work is needed. Users control permissions and confirm risky actions. Feedback will improve safety and features before wider release.
Google launched Gemini 2.5 Flash Image, a model for image generation and editing that blends images, keeps character consistency, and allows natural language edits. It’s accessible via Gemini API, Google AI Studio, and Vertex AI, offering low latency and cost efficiency. Features include multi-image fusion, world knowledge, and prompt-based editing. Developers can build, customize, and deploy apps easily. Images have a digital watermark to mark AI-generated content.
Perplexity launched Comet Plus, a $5 monthly subscription that pays publishers when users read their work or AI uses it to answer questions. It offers premium content access and compensates publishers for human visits, AI-generated answers, and agent actions, reflecting modern internet use. This model supports quality journalism, improves user experience, and integrates AI with publishing for a better web. Pro and Max subscribers get it included.
Perplexity Comet AI browser is vulnerable to indirect prompt injection, where malicious hidden instructions in web content trick the AI into accessing sensitive data. Traditional web security fails here. Mitigations include separating user commands from page content, requiring user confirmation for sensitive actions, and isolating agentic browsing. This shows the need for new security models in AI browsers as they gain power.