SIMA two Self-Improving Agent. Aligning Machine and Human Vision. Apple Requires Consent for AI. NotebookLM Creates Copyright Risk.

Show notes

The AI news for November 14th, 2025

--- This episode is sponsored by ---

Find our more about our today's sponsor Airia at airia.com.

---

Would you like to create your own AI-generated and 100% automated podcast on your chosen topic? --> Reach out to us, and we’ll make it happen.

Here are the details of the day's selected top stories:

Google’s SIMA 2 agent uses Gemini to reason and act in virtual worlds
Source: https://techcrunch.com/2025/11/13/googles-sima-2-agent-uses-gemini-to-reason-and-act-in-virtual-worlds/
Why did we choose this article?
Highlights a concrete step in agentic AI: a model that reasons, acts, and self-improves in 3D environments using Gemini. Useful for listeners tracking the frontier of embodied agents, robotics simulation, and the safety/engineering trade-offs of deployable agents.

The DeepMind study aims to align machine vision and human perception with each other.
Source: https://the-decoder.de/deepmind-studie-will-maschinenblick-und-menschliche-wahrnehmung-aufeinander-abstimmen/
Why did we choose this article?
Presents peer-reviewed research on aligning visual models with human perception, with measurable gains in robustness and generalization. Important for practitioners designing vision systems that must behave predictably for real users and regulators.

Apple’s new App Review Guidelines clamp down on apps sharing personal data with ‘third-party AI’
Source: https://techcrunch.com/2025/11/13/apples-new-app-review-guidelines-clamp-down-on-apps-sharing-personal-data-with-third-party-ai/
Why did we choose this article?
Policy change with immediate operational impact: app developers and product teams must reassess data flows to any external AI services, update privacy disclosures, and obtain explicit user consent — a practical compliance and privacy story.

You can now also infringe on copyright with Google's NotebookLM.
Source: https://the-decoder.de/man-kann-jetzt-auch-mit-googles-notebooklm-das-copyright-verletzen/
Why did we choose this article?
Flags a concrete legal and content-governance risk: a new NotebookLM capability can enable copyright violations via free-form video summaries. Practical takeaway for researchers, content creators, and platform operators to audit features and moderation controls.

Do you have any questions, comments, or suggestions for improvement? We welcome your feedback at podcast@pickert.de.