• Swindon & Wiltshire Tech
• Tuesday March 12 2024
• Author Name: Peter Davison

Princess of Wales photo controversy shows need for watermark scheme, says expert

The furore over a photograph published by the Princess of Wales shows the need for a watermark scheme for images, according to an expert from the region.

James Davenport, a professor at the University of Bath and fellow of the Swindon-headquartered BCS, The Chartered Institute for IT, said the future queen’s photo editing blunder shows the need for watermarking original images to build public trust.

The photograph, issued by Kensington Palace, showed the Princess of Wales
with her children Prince Louis, Prince George, and Princess Charlotte.

It was the first image released since she disappeared from public view two months ago following abdominal surgery.

But soon after its release the photo – apparently taken at Windsor Castle last week and released to mark Mother’s Day – was recalled by the world’s major picture agencies, which put out a ‘kill notice’ to halt further distribution.

It followed speculation by social media users that the photograph had been digitally manipulated. Blurring around Kate’s hair, distorted lines in clothing and background structures, and ‘funny fingers’ immediately recognisable to anyone who has dabbled in AI photo manipulation were quickly spotted.

Unlike portrait photographers, news photographers are banned from distributing ‘doctored’ images through the channels of reputable news agencies. Even correcting ‘red eye’ is prohibited.

Mandatory kill

“Mandatory kill. Due to an editorial issue this photo by the Prince of Wales has been withdrawn and may no longer by used in any manner. Please immediately remove it from all your online services, stop using it in any other fashion and delete it from your servers,” read the stark warning from the respected Agence France-Presse (AFP). Associated Press, the Press Association, and Reuters followed suit.

The row has sparked a discussion about the responsibility of news channels, and of social media outlets, to prevent the spread of misinformation in the form of images.

“Ever since Photoshop became widely available – people have been changing photographs with comparatively little concern,” said Professor Davenport.

“But the ability of modern technology to make systematic changes or produce completely new images – deepfakes – changes the situation.”

There is much talk about ‘watermarking AI’, but this is probably impractical, and certainly impossible to enforce – the genie is well and truly out of the bottle.

“What is really needed is watermarking ‘guaranteed originals’, which is certainly technically possible.

“However, today – the only distinction we can sensibly make is “original image” or “computer processed”.

Fear of deepfakes

Professor Davenport said watermarks might require new hardware, or a firmware change in cameras, but it is possible to sign an image to prove that it is the original, captured by a specific machine at a certain time – and possibly location.

The fundamental technology is the same as that used for digital signature of documents, with which many of us are now familiar, he added.

That is even more important as AI and Deepfake videos could play a significant part in the upcoming General Election.

Meanwhile, the royal family – whose PR strategy is ‘never complain, never explain’ – has found itself explaining, apologising, and battling even greater public speculation about the health of the Princess of Wales.

At the time of publication, the image remains on the X (Twitter) feed of The Prince and Princess of Wales.

Image: The image as it appears on the Prince and Princess of Wales’ account on the social media platform X