This is the Guardian.

It's a mental illness that will affect one in a hundred people over the course of their lives.

So schizophrenia is a term, it's quite an old term that's used to describe a collection of symptoms that some people experience, which are broadly called psychotic symptoms.

So these can include things like hearing voices,

believing things which aren't true, often paranoid thinking, but also more subtle

features like conceptual

disorganisation or social withdrawal.

Despite the devastating impact schizophrenia can have,

it isn't always straightforward to get help.

Patients can often wait, sometimes years before a definitive diagnosis is given.

There's increasing understanding in the case of psychosis and schizophrenia that the amount of time that you spend with the symptoms before being treated can have a negative impact on how will you do later on.

But I think it's fair to save it as a rule people can wait a very, very long time.

That's Matt Nore. He's researching how artificial intelligence might one day be able to improve things for patients.

These AI tools would provide incredibly valuable information to psychologists and

psychiatrists in that early diagnosis and monitoring stage.

So today we're asking how could AI change psychiatry and what the risks might be of

mixing machine learning with mental health from the guardian I'm Madeline Finley, and this is Science Weekly.

Matt Nore, you're a psychiatrist and neuroscientist at the University of Oxford and University College London

and you've got a background in computational neuroscience and you've done a study recently that kind of wraps all these

things together looking at how AI can be used to pick up on subtle differences

in speech patterns in people with schizophrenia. So how is

schizophrenia typically diagnosed? Like all psychiatric conditions it all starts from the

...