Australia is the world’s biggest Claude user, and even the ABC is now writing with it

Australia has quietly become the most Claude-obsessed country on the planet.

New data from Anthropic, the trillion-dollar company behind the Claude AI model, shows Australians use it at 6.4 times the rate our population would predict — first out of 121 countries measured, and by a wide margin.

That is up from 4.1 times just two months earlier.

We are ninth in the world by total usage, punching several weight classes above our size.

And the single most common thing we ask Claude to do?

Homework.

Almost 10 per cent of everything Australians bring to Claude is students looking for help with assigned work — course deliverables, study materials, quiz assistance, maths tutoring.

After that it is business operations, writing about yourself, promotional copy, and workplace writing.

Two states do most of the typing.

New South Wales accounts for 37 per cent of Australian usage and Victoria 31 per cent — together nearly 70 per cent of the country’s Claude habit.

Wealth doesn’t explain it.

Mining-rich Western Australia produces the most per person economically but ranks among the lowest for Claude use, which Anthropic puts down to who works there — the heaviest users are in finance, professional services and tech.

There is a distinctly Australian flavour to how we use it, too.

We lean on Claude more than other countries for medical questions, for emotional support, and for wellbeing — not just work.

And we tend to keep a hand on the wheel: 54.5 per cent of Australian use is “augmentation,” where the person stays in the loop and treats the AI as a collaborator rather than handing over the whole job. That is above the global average.

The usage even follows the clock.

Across Anthropic’s data, requests for the news peak around 7am, business emails climb through the morning, recipe requests spike 2.3 times at 6pm, and sleep advice clusters in the small hours before dawn.

The company is not content with being merely dominant.

Anthropic has opened a Sydney office, signed a memorandum of understanding with the Australian government, and its executives have met Prime Minister Anthony Albanese in Canberra.

It is now exploring building local computing capacity through Australian partners — something it says enterprises and government agencies keep asking for, especially those with data-residency rules.

Company president Daniela Amodei told Forbes Australia there is a “values overlap” between Anthropic and Australia.

“Australia has been very smart and very tech-forward,” she said.

Not everyone is convinced the overlap is so tidy.

Recent reports suggested Anthropic wanted Australia’s copyright rules watered down.

Senior government officials pushed back hard, stating the government “has ruled out a text and data mining exception” and that Australia’s copyright position has not changed.

But the clearest sign of how far this has gone is inside the country’s own newsroom.

The ABC has chosen Claude as its standard enterprise-wide AI tool.

It is running a July pilot with 100 staff dubbed “AI Champions” before expanding across the organisation.

And in the most pointed trial of all, the public broadcaster is using Claude to turn regional radio bulletins into online news articles.

The same local journalists who file the radio copy repurpose it into written stories, with a local editor and a sub-editor checking it before anything publishes.

Management framed the shift to staff bluntly.

Managing director Hugh Marks and chief people officer Deena Amorelli said “the question facing every public broadcaster is not whether they will use AI, but how they will shape the use of AI on their terms and in line with their values.”

Staff reaction has been wary, with concerns about story authenticity.

One journalist in the trial, speaking anonymously, said the AI only made minor changes to their copy — turning written numbers into digits, condensing text — and that there had been no factual errors so far.

The ABC insists a human signs off on every story, and that its distinctive journalism is something “AI cannot replicate.”

There is one catch worth watching.

The ABC has quietly narrowed its disclosure policy.

It will now flag AI use only where it “could materially affect” an audience’s understanding of the content — a looser threshold than its earlier, broader commitments to transparency.

Which means the next regional bulletin you read on the ABC may have been written with Claude.

You just might not be told.

———

The Glass and AI

A word on how we use it, since this story is about everyone else.

The Glass uses AI, including Claude, in its newsroom. We use it to help research, verify, sub-edit and format stories, and to draft social copy.

We do not publish anything a human editor has not checked, corrected and approved. Every fact is verified against original sources. Every story is signed off by a person.

We will not pretend otherwise, or bury it in a policy you have to go looking for.

If AI shaped a story in a way that would change how you read it, we will tell you — not only when we decide it “materially” matters.

That is the whole point of a masthead called The Glass.