AI

The conversational mode for Anthropic’s AI chatbot is now becoming available for users of its mobile apps, “gradually.”

Mary Hubbard, the executive director of WordPress.org, said the group will work on “accelerating and coordinating artificial intelligence projects across the WordPress ecosystem.” The team will also maintain a public roadmap of its AI plans and plugins, which it will share on its new page for “Core AI.”

The Atlantic has a fascinating deep dive into khipus — long cords that the Inca tied knots into to preserve information. Few know how to read the knots, which are hundreds of years old and fragile. But researchers are slowly learning to understand them:

Amazon engineers related their experience creating software to The New York Times:

Unionized workers at Politico allege the company violated its contract by using AI-generated content in a live blog — and are escalating their complaint to arbitration this summer. The union says the AI-generated summaries contained factual errors and language that’s off-limits for writers. The result of arbitration could set an important precedent for an industry that’s seen countless AI disputes in recent years.

The Justice Department is investigating whether Google crafted its agreement to skirt regulatory scrutiny, Bloomberg reports. The deal brought Character.AI’s co-founders back to Google and didn’t technically involve an exchange of shares, though investors were set to receive a payout, The Verge previously reported. Google spokesperson Peter Schottenfels told Bloomberg that Google is “always happy to answer any questions from regulators,” and added that Character.AI remains separate, with no ownership stake by Google.

Surge AI is being sued over worker misclassification in California. The class-action suit was brought by the same firm that filed a similar suit against Surge’s larger rival, Scale AI, late last year. The complaint alleges that Surge exerts an extremely high level of control over how and when its contractors work and also accuses the company of failing to pay wages for the copious training and testing contractors must perform before being allowed to start on a project.

The first lady has announced that the new audiobook version of her self-titled (and suspiciously robotic) memoir was entirely narrated using an AI clone of her own voice created by ElevenLabs.

The excellent Data & Society pulled together a report on the impact of generative AI on scamming. (No surprise: not good.) Its conclusion makes an interesting comparison to the once-pressing scourge of spam:

The Verge team has a few guesses about the first hardware release from the OpenAI and io combination. Let us know yours in the comments.

The last question during the AI fireside chat at I/O 2025 was an invitation to make headlines, and the Google co-founder did his best, saying... something about reality and our existence. Listen in for yourself.

Brin showed up to crash Google DeepMind CEO Demis Hassabis’ fireside chat at I/O 25, where he laid out what he does all day when asked by host Alex Kantrowitz.

You can now have NotebookLM make you Audio Overviews that are short (around 5 minutes) and long (around 20 minutes) in addition to the default length of around 10 minutes.