AI News 26th November 2025

AI News Wrap-Up: 26th November 2025

🧒 EU MPs push 16+ default for chatbots and social apps

The European Parliament backed a non-binding call for a harmonised digital minimum age of 16 to access social media, video platforms, and AI companions.
Teens aged 13–16 could still join with parental consent, while under-13s would be blocked-at least on paper.
Lawmakers also float bans on addictive design, manipulative ads, and the ability to block non-compliant sites. Spicy-but not law yet.
🔗 Read more

🧠 USPTO says AI can help invent - but only humans get the patent

New guidance treats generative AI like lab gear-useful, powerful, still just a tool.
No special inventorship tests for AI-assisted work, and yes: the inventor on the patent must be a person, not a model with vibes.
It also drops an earlier joint-inventor framing that, honestly, only muddied things.
🔗 Read more

🖥️ Mexico unveils ‘Coatlicue’ supercomputer to turbocharge AI

The plan: build Latin America’s most powerful machine for AI and data crunching, nicknamed after a Mexica earth goddess. Bold choice; cool name.
Target performance: 314 petaflops-about seven times the region’s current champ in Brazil-with construction starting next year.
Officials say the country’s top box today peaks at 2.3 petaflops, so… big glow-up incoming.
🔗 Read more

💼 HP to cut up to 6,000 roles as it leans into AI

HP says reshaping for AI will speed product development and boost productivity-or so it hopes.
The plan targets $1bn in annual savings by 2028, with restructuring costs near $650m, touching product, ops, and support teams.
Markets didn’t exactly cheer, but the company insists AI PCs are a bright spot. Careful pivot-or hard swerve?
🔗 Read more

🎵 Warner Music strikes licensing deal with AI song generator Suno

After suing Suno, Warner flipped to partnership mode-artists can opt in, get paid, and see new licensed models roll out.
Downloads get tighter rules-paid tiers, caps, extra fees-while Suno also picked up Songkick. From Napster-era vibes to something more Spotify-ish.
Not everyone’s hugging it out yet-other label fights continue-but this is a notable détente.
🔗 Read more

🔐 ‘HashJack’ shows how URL fragments can hijack AI browsers

Researchers tucked malicious prompts after the # in URLs to steer AI assistants toward phishing, data leaks… the works.
Perplexity’s Comet looked especially exposed; Microsoft and Perplexity pushed fixes, while Google called the behavior in-bounds-which is, uh, a take.
Some tools resisted the trick, but the bigger lesson is simple: links can talk to your AI even when pages don’t.
🔗 Read more

Yesterday's AI News: 25th November 2025

Find the Latest AI at the Official AI Assistant Store

About Us

Back to blog