Category: About Ai

Amazon has rolled out a new AI-powered feature in the Kindle app called Ask This Book, allowing readers to ask questions about the story and receive spoiler-free answers based only on the pages they’ve already read.

Key highlights:

• Answers are generated strictly from the portion of the book you’ve read
• Accessible via the book menu or by highlighting any line
• Responses can’t be copied or shared and are limited to book owners or renters
• The feature is always on — authors and publishers can’t opt out
• Currently live in the U.S. for thousands of English-language titles on iOS
• Kindle devices and Android support are planned, but no confirmed timeline yet

This follows Amazon’s earlier AI Recaps feature and arrives shortly after the company removed some inaccurate AI summaries from Prime Video.

🔍 Open questions:

• How will AI reshape deep reading experiences?
• Where should the line be drawn on author and publisher control in the age of AI?

Via mindstream reporting

In response to Google’s AI upgrades, OpenAI CEO Sam Altman announced an internal “code red” to accelerate improvements to ChatGPT. Key points:

• Short-term boost: New reasoning model Shallotpeat reportedly surpasses Gemini 3 on benchmarks.
• Long-term vision: Larger model upgrade Garlic (GPT-5.2/5.5-type) targeting 2026, aiming to fix pre-training limitations.
• Strategic focus: Delaying advertising and AI agent projects to prioritize user experience, personalization, and image generation.

The AI race is heating up — just as OpenAI once sparked Google’s emergency push in 2022, now the tables have turned. With rivals like Gemini 3, Opus 4.5, and open-source initiatives, the competition for AI leadership is more intense than ever.

Via therundown reporting

Meta has expanded its AI partnerships, announcing that its chatbot will now use information from CNN, Fox News, USA Today and People Inc.

It has also signed deals with The Daily Caller, The Washington Examiner and France’s Le Monde.

This comes as publishers continue to challenge AI firms over how their work is used.

Meta says these agreements will help its AI give faster, more reliable information from a wider mix of sources.

Here’s what you should know:

• AI companies are turning to licensing as legal disputes grow.
• Publishers are using lawsuits to set boundaries on how AI models use their work.
• Tech platforms are rethinking how they work with news, focusing on clearer agreements.

A new playbook?

The shift toward licensing follows Meta’s decision to move away from traditional news deals, close Facebook’s News tab and remove news from its platforms in Canada after a law required payment to publishers.

Other AI companies are following a similar path.

OpenAI has licensing deals with outlets including The Wall Street Journal, the Financial Times, Future and Vox Media, while also facing its own lawsuit from The New York Times.

Every time AI companies announce a licensing deal, a lawyer somewhere buys a nicer coffee. – MV

Via mindstream reporting

The New York Times has filed a new lawsuit against AI search startup Perplexity, accusing the company of using its journalism without permission.

Other outlets, including the Chicago Tribune, have recently made similar claims.

The Times says Perplexity’s tools pull information from news sites, sometimes from behind paywalls, and turn it into answers for users through its RAG-based chatbots and Comet browser assistant.

It also claims some responses copy its articles too closely or attribute inaccurate information to the outlet.

This legal action sits within a wider industry push.

Many publishers are negotiating licensing deals with AI companies, but lawsuits have become a way to pressure firms into paying for news content.

Perplexity has introduced revenue-sharing programmes and signed a licensing deal with Getty Images, but the Times argues its content is still being used without agreement.

Perplexity says publishers have challenged new technologies for decades, but the case adds to growing scrutiny.

News Corp, Britannica, Nikkei, Reddit, Wired and Forbes have all criticised the company’s data practices, and Cloudflare recently confirmed that Perplexity scraped sites that had blocked it.

In short:

• The Times claims Perplexity uses its journalism without permission, including paywalled material.
• The lawsuit is part of a wider push for licensing and compensation from AI firms.
• Courts are still deciding how copyright rules apply to AI training and AI-generated answers.

Paywalls are not having it

The Times is also suing OpenAI and Microsoft over how its articles were used for AI training.

Courts are still defining what “fair use” means in these cases, though a recent settlement in a similar lawsuit against Anthropic suggests that licensed and unlicensed data may be treated very differently.

The Times wants damages and a court order stopping Perplexity from using its work.

At the same time, it continues to sign paid licensing deals with other AI firms, reflecting how publishers are trying to find sustainable ways to work with AI.

AI companies beefing with newspapers feels like Season 4 of a show nobody wrote. – MG

Via mindstream reporting