AI Insights

Claude to bring US voting information into conversations

Anthropic outlined its midterm election preparations on October 14. Later this month, it plans to include state-specific answers based on official voting information collected by Democracy Works.

Create slides in ChatGPT and edit them together

Slides creates presentations from conversations or files and supports direct editing of text, images, shapes and charts. It is rolling out to eligible Pro, Business and Enterprise users with Space access.

OpenAI offers sponsored Rosalind access to US university labs

Eligible labs can receive up to 30 Business Standard seats at no cost for up to 12 months after approval. GPT-Rosalind requires trusted-access approval, and API access is not included.

Cursor cloud agents can build and share interactive canvases

Cloud agents can now create interactive canvases and share them with teammates through a link. An API for creating and managing cloud environments is also available in beta.

GitHub makes AI Scan generally available for pull requests

AI Scan checks changes for vulnerabilities, including languages and frameworks not covered by CodeQL. Its findings are advisory and do not themselves block merges.

Copilot can require current company policies at startup

GitHub added a setting that requires supported Copilot clients to retrieve current managed settings at startup. If the policy cannot be confirmed, the client stops rather than using stale settings.

Kling 4.0 launches with single-pass video up to 30 seconds

Kuaishou announced that Kling 4.0 launched on October 15. It supports up to 30 seconds of video, up to 15 reference assets and 10 keyframes to guide characters and scene progression.

Reverie makes continuity the default for new clips

New clips now default to Continuation, building on the previous frame. Users can choose Hard Cut when they want to change location or time.

Reverie adds intercut clips for consistent dialogue scenes

Reverie launched Intercut Clips, which references both characters in a conversation. The company describes it as a way to preserve faces and settings while cutting between reactions and replies.

YouTube Music adds player labels for AI-generated tracks

YouTube is rolling out labels for tracks created primarily with generative AI. Listeners can open more information from the player, and the labels will also expand to the main YouTube app.

Google launches Deep Review to check complex documents

Google expanded its pre-submission paper review system to broader document verification. At launch, it accepts PDFs and plain text and offers two complimentary reviews per week.

Google links AI design to feedback from laboratory experiments

Google Research described work connecting AI agents to laboratory robots. It reported infrastructure that tracks reagents and equipment, along with prototypes that feed experimental results into subsequent molecular design.

Google expands experimental access to Earth AI agents

Google published research on agents that combine satellite, weather and population information. It is expanding early access through Ask Google Earth and an agent-to-agent API.

ElevenLabs reports renewed security and AI governance certifications

On October 16, ElevenLabs reported renewals of its information security and AI governance certifications. It also listed added attestations and certifications, including SOC 2 Type I for ElevenProductions.

Cursor Origin connects to over 20 development tools

Origin's marketplace now includes integrations with services such as Cloudflare, Linear and Sentry. External developer teams can submit apps for review and listing.