AI can easily alter your social media posts, and there's little protection, expert warns

Little can be done to protect social media photos from manipulation by Artificial Intelligence, warns a University of Guelph cybersecurity expert.

This comes after two Guelph, Ont. boys were caught using AI to alter several classmates' social media photos, turning them into explicit images.

The incident may leave parents wondering how they can best protect their child's photos from being inappropriately modified by deepfake technology.  

Ali Dehghantanha, professor in cybersecurity and threat intelligence at the University of Guelph, says AI deepfake technology is evolving at a fast pace, making it challenging to protect any photos posted online. 

"It's very difficult," he said. "With all the advancement in AI technologies, it's quite trivial for the AI to be able to create fake content."

'Do not share any images online,' says expert

One method that can be used to safeguard your child's photos is to add a digital watermark. This allows a user to track a photo they posted. Dehghantanha says this method can be easily by-passed.

"The deepfake can easily remove those watermarks," he said.

Dehghantanha says the only way to ensure your social media photos are safe is to stop posting on social media and delete all of your photos. 

"If you feel uncomfortable, do not share any images online, that's all you can really do," he said.

Deleting your social media photos still leaves them open to being manipulated as the photos often remain online.

Dehghantanha recommends contacting social media networks to ensure your pictures have been deleted. 

"You can always ask them to send a report of what they know about you so you can even double check that what you wanted deleted is actually deleted," he said. 

Dehghantanha stressed contacting social media networks does not guarantee your photos are safe. 

"It's impossible to just remove everything about yourself online," he said.

Changing definition of privacy online

Dehghantanha added that he believes photos and videos may be seen as untrustworthy sources of information online in the future. 

"Our reliance on videos or images as a reliable source would change significantly," he said. 

While it is possible for social media networks to determine if a photo has been manipulated by AI, Dehghantanha is skeptical whether they will actively try to find and remove AI manipulated photos.  

"The technology is there, but there isn't an intention, law or regulation," he said.

Going forward, Dehghantanha believes people will accept the risk that their social media photos could be manipulated by AI.

"The definition of privacy will change, some of the things we consider private today won't be considered private anymore because there's no way to protect us," he said. 

Mason Melitzer, member of the youth and community services unit with the Guelph Police Service, stressed that it's the parents who need to take responsibility in monitoring their child's online activity.  

"It's very important to make sure we know what kids are doing online, what they're accessing, who they're communicating with," he said.

"If we're wanting to give are youth access to all of this, we need to be monitoring what they're doing."