AI ‘deepfake’ applicants are flooding university intakes

UK universities are increasingly facing the threat of ‘deepfake’ applications from prospective students, as AI software risks deceiving interview systems and automating applications.

Enroly, a software programme used by universities to streamline applications, noted that a small proportion of students have attempted to use deepfake technology in order to manipulate and submit video interviews. 

‘Deepfake’ technology takes AI-generated images and audio in place of real-life alternatives, allowing students to submit applications with AI-altered accents or to hide the fact that someone else is responding to their interview questions. 

This has raised questions regarding fraud within university applications, especially with the ongoing advancement of AI technology.  

It’s like something out of a spy film […] they’re incredibly hard to detect

Phoebe O’Donnell, Enroly’s Head of Services

Despite only 0.15% of applications being detected as deepfakes, Enroly has warned of “the future of fraud”, highlighting the increasing concern of attempts to submit fraudulent, AI-vetted applications.  

Phoebe O’Donnell, Enroly’s Head of Services, said that deepfakes are “the stuff of nightmares for interview assessors – fake faces layered over real ones, complete with expressions and movements… It’s like something out of a spy film. And yes, they’re incredibly hard to detect.” 

Enroly has said that multiple processes are in place to identify fraud, including passport matching and facial recognition.  

Universities are coming under increasing pressure to […] find solutions to the ever-growing usage of AI in all aspects of academia

Universities may also find it appropriate to implement their own methods for international student applications. Currently, some British universities use online interviews and questionnaires to screen international applications before issuing the Conformation Acceptance for Studies certificate – essential for the Home Office to issue a visa.  

These automated interviews require students to record answers to a random selection of questions via an online platform, which are eventually viewed by admissions staff who can request live interviews if they flag interviews for being suspicious.  

Even though deepfakes are not limited to university applications, universities are coming under increasing pressure to identify issues and find solutions to the ever-growing usage of AI in all aspects of academia.