Amid the proliferation of hoaxes perpetrated by generative artificial intelligence (AI) or deepfakes, consumers appear conscious of the growing trend, per an online identity consumer study.
Deepfakes are media produced using digital manipulations via AI, aimed at the creation of believable impersonations or events.
The study, which was conducted in April this year, saw 8,055 participants at equal distribution from across Mexico, Singapore, the United Kingdom, and the United States.
In it, it’s revealed that 67 percent of consumers have at least heard of deepfakes—the majority of which, or 87 percent, consists of Singaporeans, according to the survey released by global market research consultancy firm Censuswide and identity verification company Jumio.
Despite advancing to a “level of sophistication that prevents detection by the naked eye,” the study showed that 52 percent are nevertheless confident in their ability to detect a deepfake video.
Jumio CTO Stuart Wells, however, believes that deepfakes are “getting exponentially better all the time and are becoming increasingly difficult to detect without the aid of AI.”
Wells also gives credence to the notion of enterprises increasingly adopting AI technologies to offset the implications of deepfakes with their consumers. In particular, companies within the crypto and payments spaces.
As for the consumers themselves, the Jumio exec is encouraging the public to maintain a healthy level of skepticism when facing provocative provocative “images, videos and audio,” which is often unmasked with some “quick research.”
In light of the participants’ awareness of deepfakes, the study also unveiled that 57 percent believe that online identity theft will become more prevalent arising from the availability and advancement in AI, with Singaporeans displaying the highest level of cognizance of its repercussions at 73 percent.
Jumio Chief of Digital Identity, Philipp Pointner, said that, while education cannot put a stop to the scheme, he recommended the use of multi-modal and biometric-based authentication systems in detecting deepfakes and preventing identity theft.