Did artificial intelligence applications reveal mimicked voices؟
Artificial Intelligence Applications is increasingly infiltrating our lives, exerting control over various professions and domains. One application of AI, known as deepfakes, involves generating voices, videos, and images that appear to originate from humans.
“Engineer Tech” came across a study conducted by “College London University” that demonstrated that over a quarter of mimicked voices using AI applications, specifically the deepfake technology, successfully deceived even the most attentive and focused listeners.
According to the university researchers, during the study, they were only able to detect 73% of the mimicked voices, even after training more than 500 individuals to differentiate between mimicked and authentic voices. This means that the detection failed by a margin of 27%.
The researchers found similar results across two languages, Mandarin Chinese and English. Although Mandarin speakers indicated that the distinction was based on rhythm, fluency of speech, and speed, the research team discovered that English speakers differentiated between a mimicked voice and a human voice through the rhythm of breathing, utilizing AI applications.
It’s worth noting that a major technology company confirmed its intention to launch a new application in the upcoming days. This application can replicate genuine human voices by utilizing the real voice for a mere three seconds.
As a result, experts in AI applications believe that we are on the verge of further similar experiments, advancing the concept of artificial intelligence and its utilization in diverse sectors such as banking, healthcare, educational sessions, retail, and even the development of websites.
A warning came from the team at the British university, stating that it is “possible to create a near-realistic replica of anyone’s voice by utilizing a recording of a few words spoken in their own voice. This is a result of technological advancements.”
The team further explained that the research results might not fully reflect the actual reality since the participants in the study – even those who were not trained – were aware that they were part of an experiment. Therefore, they were more focused on detecting the mimicked voices.
Concerns have been raised regarding recent advancements in artificial intelligence technology, particularly in the development of deepfake technology, which could lead to widespread fraud through voice impersonation. The ability to accurately mimic authentic voices through AI applications has raised worries.
It’s worth mentioning that these accurately mimicked voices have been exploited to deceive individuals and extort money from them. Scammers have used these voices to falsely claim to be friends or business partners, convincing victims to transfer money to them.
Currently, most efforts to combat forgery are centered around the use of AI applications. These applications performed at a similar level to the study participants in the university experiment, yet they were more successful compared to those who were unaware they were participating in a voice detection experiment.
Let us not forget that artificial intelligence applications help humans quickly and instantly understand their surroundings and harmonize with those around them. This stems from the demonstrated capability of technological solutions to respond to and interact with humans according to their immediate needs.
