China's AI companies are running short on engineers, so they're now scouting talent as young as 13 with camps, mentorships, and even guaranteed jobs straight out of high school.

Gemini Notebook’s new Collections help users organize growing research libraries as Google expands the former NotebookLM across Gemini and Search, although the feature remains more limited than a proper folder system.

Samsung wants to build humanoid robots, but its new RX division is starting with practical factory machines while consumer models and robot butlers remain distant ambitions without a launch timeline.

You probably won't notice the next Model Context Protocol update, but it could make AI assistants much better at connecting to the apps and services you use every day.

In a decisive move to protect viewer feeds and advertiser dollars, YouTube has updated its monetization guidelines to strip cash flow from low-quality, AI-generated material. The streaming giant is refining the rules governing its lucrative YouTube Partner Program (YPP) to specifically target “inauthentic content” designed solely for low-effort content farming. Rather than banning artificial intelligence […]

A Seoul National University team built an AI that scores self-driving cars' safety choices in real time, and it just landed a rare CVPR highlight paper nod.

A cybersecurity researcher poisoned an open weight AI model for under $100 in about an hour, exposing how easily these increasingly popular systems can be secretly compromised without detection.

Every app wants to become an AI app, even when users keep dismissing the invitation. Somehow, software that promises to understand us still can’t remember when we’ve said no.

Google used live stadium capture, and combined it with archival material to bring Pele's legendary goal to life. The Veo video engine, Gemini Omni, and Nano Banana Pro AI tools tagged alongside in the journey to revive a lost marvel.

Anti-surveillance clothing is becoming affordable enough for ordinary wardrobes, but its patterns only confuse certain computer vision systems. For now, it works better as a privacy statement than reliable protection.

Moonshot’s 2.8-trillion-parameter Kimi K3 comes close to Fable 5 and GPT 5.6 Sol in several benchmarks, while remaining open weight.

Meta is adding parental alerts and human review to sensitive AI conversations involving teens who may be at risk of self-harm.

OpenAI and Work Louder built a $230 mini keyboard that gives Codex users physical shortcuts, live agent status lights, and a dial for adjusting how hard the AI thinks.

Google disputes a child-safety assessment of AI Overviews and AI Mode, arguing that its testing was unrealistic. The report documents troubling responses and questions whether parents have enough control over Search.

Forget Google Translate, this AI tool invents languages nobody has ever spoken, sounds, grammar, and all.

Reelful is an AI app that builds finished reels from your photos and clips, complete with voiceovers, captions, and music.

South Korea plans to give every citizen free, unlimited access to a domestically powered AI chatbot, with a September beta and a separate agent designed to help people navigate government services.

Samsung says refusing AI training in Samsung Health will only delete data collected for AI development, leaving existing health records untouched.

Spotify is rolling out a conversational AI feature that can play requested tracks, answer music questions, and refine selections through follow-up prompts.

Researchers tested Gemini 2.5 Pro on parent-child interaction videos and found AI is reliable at observation but falls short when making expert-level judgments about child communication.