File photo: The app can even send alerts to parents, but the aim is to get the youngsters to curb their own behaviour. Picture: Max Pixel

London - It's a dilemma for every modern parent – how to keep children safe on social media without monitoring their every post.

Now the developers of a new smartphone app claim to have come up with the answer – a friendly "bot" that warns youngsters to think again before they send naked photos, explicit messages or even personal details online.

Oyoty uses artificial intelligence which detects when a picture has too much flesh on show, and asks the child to think again.

The app can even send alerts to parents, but the aim is to get the youngsters to curb their own behaviour.

Social media giants such as Facebook and Snapchat have been criticised by child-safety campaigners for failing to tackle the problem.

Computer experts in Switzerland have spent years programming the Oyoty app so it can immediately distinguish harmful material.

It also guards against abusive and bullying language and revealing sensitive information such as phone numbers.

The app is designed to run in the background of a device, sweeping for dangerous posts on Facebook, Instagram and Twitter.

If it spots something suspicious, a "chat bot" engages with the youngster. In one example, the app sends a message to a child, asking: "There is a lot of skin showing in this picture. I wonder if you might need to stop and think about sharing this?"

It then asks: "Would your parents/carer think it was a good idea, or will you be embarrassed if they happened to see it?"

It then guides the youngster through how to delete the image. App developer Deepak Tewari said: "This is not about telling children what to do but about educating them and helping them to make their own informed choices.

"Very young children are going online in increasing numbers. The industry is not doing enough to recognise it as a problem. Studies suggest children aged eight, nine and ten are spending ten hours a day on their platforms unsupervised."

He said none of the harmful images or posts is ever retained by the app – and he hopes that the technology will eventually come automatically installed on phones used by children.

Mail On Sunday