AI security experts and researchers in Aberdeen are providing updated techniques to identify AI-generated images, videos, and scams [1, 2].
This shift is necessary because deepfakes have become increasingly convincing, creating significant risks for misinformation, security breaches, and sophisticated phishing attacks [2, 5].
For years, the public relied on specific visual glitches to identify synthetic media. However, experts said that these old clues are no longer sufficient in 2026 [2]. The rapid evolution of generative AI has smoothed over the errors that once made deepfakes easy to spot, such as distorted fingers or unnatural blinking, making the media nearly indistinguishable from reality.
Research conducted in Aberdeen, Scotland, focuses on training people to recognize these newer, more subtle markers of AI generation [1, 2]. These efforts aim to counteract the rise of AI-phishing scams, which experts said have become harder to detect [4].
Earlier guidance from March 2024 provided basic tips for spotting deepfakes, but those methods have quickly become obsolete [3]. The current landscape requires a more critical approach to media consumption, as AI tools can now mimic human speech and movement with high precision [5].
Security specialists said that the goal is to move beyond looking for a single "tell" and instead analyze the context and source of the media. Because AI is now everywhere, the burden of verification has shifted toward a more rigorous process of cross-referencing information [2].
“Old clues are no longer sufficient by 2026”
The diminishing reliability of visual artifacts in AI media marks a transition from a 'detection' era to a 'verification' era. As generative AI eliminates the obvious technical flaws of the past, the ability to distinguish truth from fabrication will rely less on spotting glitches and more on the provenance of the data and the credibility of the source.
