Experts say LLMs aren’t intelligent and never will be

Researchers behind a widely cited Nature commentary put it bluntly: "Language is primarily a tool for communication rather than thought."

Turns out the secret to jailbreaking AI is poetry

Across all the tests, poetic jailbreaks succeeded 62 percent of the time.

ChatGPT turns 3, and it has changed how people look things up

OpenAI CEO Sam Altman predicted that “someone is going to lose a phenomenal amount of money in AI.”

AI won’t admit it, but it probably is biased

Reminds you of old AI days when “professor” always meant an old man and “student” always meant a young woman.

How computer vision learns to read the streets

AI is revolutionizing how we detect graffiti, turning maintenance into a precision operation.

US Patent Office says AI can help invent, but can’t take credit

AI can brainstorm, analyze, and suggest ideas, but only a real human has to be the one who actually conceives the...

AI can crack jokes but still doesn’t get your puns

While AI can produce a perfectly structured joke, it often has absolutely no clue why that joke is supposed to be...

Are we about to pop the humanoid robot bubble?

Engineers are still struggling with basics like giving robots hands that don’t crush everything they touch, or making bodies that don’t...

Anthropic to face Congress over Claude’s China use

Anthropic detected suspicious activity in mid-September and uncovered a highly sophisticated espionage campaign.

OpenAI lets the problematic AI teddy bear back in

It will now run on GPT-5.1 Thinking and GPT-5.1 Instant, instead of GPT-4o model.

OpenAI denies responsibility in teen wrongful death lawsuit

Adam used ChatGPT for about nine months, during which the chatbot allegedly encouraged him to seek professional help more than 100...

Regular ChatGPT users dodged a bullet in latest AI security breach

OpenAI's analytics partner, Mixpanel, hit a speed bump with a security incident that exposed developer info—names, emails, and locations—for a select...

Character.ai trades chatbots for stories in under-18 safety measure

This is to reduces the risk of unhealthy attachment while still letting teens enjoy the creative side of the platform.

ChatGPT now lets you use text and voice mode in the same screen

Instead of bouncing between modes, voice conversations will happen right inside your normal chat thread.

Judge tells OpenAI to drop “Cameo” from Sora (for now)

Cameo lets users generate AI videos of anyone you want: a celebrity, a creator, or even your friend by uploading their...

Jony Ive and OpenAI screenless AI gadget wants to replace your phone

Soon you will be carrying a tiny, screenless AI device in your pocket, quietly judging your life choices, but in a...

Anthropic’s new Opus 4.5 comes with Chrome and Excel integrations

It’s the first model ever to score over 80% on SWE-Bench Verified

OpenAI locks down after alleged protester threat

Employees were told to stay inside, avoid wearing OpenAI-branded clothing, and even remove their badges when leaving the building.

ChatGPT’s ‘toxic positivity’ blamed for suicides in shocking lawsuits

OpenAI says it’s improving safety features, rerouting sensitive chats to newer, less yes-man-y models like GPT-5.

Perplexity’s Comet AI browser is now available on Android

You can ask Perplexity questions about specific tabs by mentioning them, or even use voice mode to quiz the browser about...