AMEINAH ALZEYOUDI (ABU DHABI)
The rapid development of artificial intelligence is making it increasingly difficult for social media users to distinguish between authentic and AI-generated content, cybersecurity experts have warned, as realistic images, videos, voices and manipulated media become more accessible.
According to Maher Yamout, Lead Security Researcher of Kaspersky's Global Research & Analysis Team, the widespread availability of AI technologies has made it simpler to produce extremely life-like content, posing significant difficulties for people attempting to discern what is authentic.
"We live in an era where artificial intelligence and tools that use it are widely accessible,” he said. "It is now possible to create or modify text, images, videos and voice recordings with amazing realism, which facilitates the rapid dissemination of false information on social media and messaging platforms."
According to him, developments in deepfake technology are making it harder to tell the difference between true and modified audio and visual content.
"Voice and visual forgeries can be nearly indistinguishable from authentic messages, as traditional cues used to spot fraud disappear,” he said. Criminals can also make use of perseverance and urgency to make false information seem more realistic.
Unconfirmed AI-generated content might cause uncertainty and erode public confidence in reliable information sources, indicating that the implications go beyond specific schemes.
"It can cause confusion and misunderstanding about important issues when people can't tell the difference between real and fake," he stated.
Yamout also emphasised the possible repercussions of false information about delicate topics like healthcare. He claimed that false information about serious illnesses or medical procedures could cause people to make poor choices.
AI-generated content can also be used to escalate disputes, incite fear, disseminate misleading information and trick people into thinking that fake content originates from a trustworthy source.
Users should be on the lookout for tiny visual and auditory cues that could indicate AI manipulation, according to Wael Jaber, CSO of Cyber Knight.
"While AI-generated content is becoming more realistic, there are still several indicators users can look for," Jaber stated, citing objects that seem to merge or change shape unexpectedly, distorted hands and fingers, uneven lighting and shadows, unnatural facial expressions and irregular reflections.
Jaber advised viewers to be aware of lip-sync issues, odd eye movements, erratic audio quality, and abrupt visual defects in videos.
Additionally, context might serve as a crucial warning indicator. According to him, information that seems dramatic, intensely emotional or "too extraordinary to be true" ought to be examined more closely before being accepted or disseminated.
Jaber advised people on social media to simply "pause before you share." This entails verifying the original source, contrasting the assertion with reporting from reputable news outlets and, when feasible, employing reverse image searches.
Additionally, users should look for provenance signs, watermarks or metadata when accessible, and check questionable content for visual or auditory abnormalities.
"Verification is becoming as important as content consumption itself in today's digital environment. "Spending a few more minutes verifying authenticity can stop thousands of people from being misinformed."