Can you spot an AI image? Quiz shows how difficult identifying deepfakes has become.

The Growing Challenge of Identifying Synthetic Media

Can you spot an AI image - A new study conducted by a firm specializing in identity checks revealed that US citizens possess a level of precision comparable to flipping a coin when separating genuine photos from computer-generated counterparts. This surprising statistic highlights the rapid evolution of digital manipulation technology, which has made traditional visual cues increasingly unreliable for the average viewer. To understand the real-world implications of this data, Kristine Lazar, who investigates consumer issues for CBS Los Angeles, challenged this statistic by taking a recognition test. She achieved a low score of three correct answers out of twelve possibilities prior to receiving instruction on key indicators.

Lazar, a seasoned professional in her field, expressed her surprise at the outcome of the assessment. Despite her extensive experience, the nuances of artificial intelligence had caught her off guard. She remarked on her performance, stating:

"I've been a consumer investigator for nearly a decade, but I got an F when I tried to identify images and videos made with AI."

The Veriff Assessment

The quiz was developed by Veriff, a prominent online platform dedicated to identity verification services. The assessment presented participants with a series of side-by-side comparisons featuring both authentic photographs and synthetic media. Initially, Lazar felt quite certain about her selections, trusting her instincts to guide her through the exercise. However, as she reviewed the answers, she realized that many of her initial assumptions were fundamentally incorrect. One particular image confused her significantly, as she had labeled it as artificial when it was actually a genuine photograph.

Reflecting on that specific moment, Lazar noted her misjudgment: "I thought he was the fakest-looking one there," she recalled, admitting her error after the reveal. Ultimately, her final tally stood at only three correct identifications, representing a mere thirty-three percent accuracy rate. This result aligns closely with broader trends observed by the verification company.

Raul Liive, who serves as the product director at Veriff, explained that their internal data mirrors the public's struggle. He emphasized that the difficulty is not unique to Lazar but is a widespread phenomenon. Liive stated:

"We are seeing that people in America especially are close to, like, it's a coin flip essentially,