Are Your Photos AI-Generated? 2 Key Techniques to Spot Deepfake Flaws

Based on discussions with experts invited by a VentureBeat reporter, this article explores two main methods for addressing image authenticity issues. These techniques can help us meticulously analyze subtle anomalies in images and verify the true origin of the content.

Technique 1: Check if Specific Image Details Differ from Reality

If, upon observing an image, you notice unnatural gestures, mouth movements, and blinking in people, or perceive abrupt inconsistencies in skin texture, background, and lighting, it is highly likely an AI-generated product. Vekiarides, CEO of Attestiv, a company specializing in detecting fake AI-generated images, advises careful observation of details such as shadows and objects, including fingers and teeth. Similarly, Matthieu Rouif, CEO of Photoroom photo editing tool, shares another perspective: examining facial symmetry. He points out that AI-generated faces often appear perfectly symmetrical, which is uncommon for human faces.

Technique 2: Verify Image Sources and Maintain Skepticism

Matthieu Rouif points out that whenever users encounter suspicious images online, they should remain skeptical and verify the content repeatedly, especially for advertisements that might involve personal data or payments.

Grobman, CTO of cybersecurity company McAfee, further reinforces this, encouraging people to always remain skeptical when encountering images and actively verify the source, especially when advertisements request personal information or payments. This is to prevent potential fraud and protect personal privacy.

Currently, several companies are dedicated to developing detection technologies to counter the threat of disinformation. For example, Google and ElevenLabs have developed technologies capable of verifying content authenticity. Concurrently, McAfee has introduced a technology that identifies AI-generated voices, with Grobman adding that its accuracy rate reaches up to 90%. This effectively detects and prevents malicious AI-generated content, further safeguarding user rights. This demonstrates the advancements and efforts of technology in ensuring cybersecurity and protecting user interests.