Child safety experts have warned internet users to stay clear of a website that purports to protect children by monitoring chat rooms for suspicious adults.

The site, known as chatnannies, claims to be "the first and only service to protect your children online and in chatrooms from paedophiles, perverts and other undesirables".

Set up by software developer Jim Wightman, the website asks people to sign up as volunteer chatroom monitors. Mr Wightman also claims to have developed artificial intelligence (AI) software that complements the volunteers.

"Nanniebots behave like humans, sound like humans, and report to ChatNannies as if they were human, but with one massive difference - they never sleep," the website says.

"Nanniebots self replicate so, as you read this, there are already thousands of them watching over chatrooms."

AI experts have questioned this, saying it would represent a massive and unlikely advance on current technology.

"If true, this would make the software 10 years ahead of what is currently available," said Henry Thompson, reader in AI at Edinburgh University.

Leading child safety experts have also advised internet users not to get involved. John Carr, the internet consultant for children's charity NCH, said: "This is vigilantism and amateur dabbling and there is no space for outfits like this. How can you vet the people you sign up to monitor chatrooms?

"It is also not the first site; there is the Stopitnow site supported by organisations such as the NSPCC, NCH and Barnado's."

The Internet Watch Foundation also warned against the site. In a statement it said: "We do not advocate monitoring of websites by arbitrary individuals."

The Association of Chief Police Officers stated that there were already a number of projects in place to protect children. "Protecting children from sexual abuse is a difficult and often complex task," it said.

"There are a number of joint initiatives to help link education and law enforcement ensuring that children are provided with the maximum support to stay safe online."

Mr Wightman defended the site, saying: "Once we are able to talk properly with these child protection agencies then perhaps they can raise these fears with us directly and allow us to allay them before they approach the press.

"We understand the uncertainty and nervousness about this new technology but we can only allay fears by being as open as possible, and for agencies to be as open as possible with us.

"We are confident that, once these agencies are fully informed as to our technology, we will be in a very strong position."