New Delhi: People are not only falling for images and videos created using artificial intelligence, but also tend to trust AI-generated faces more than real ones, say researchers from multiple universities in a new study.
According to Lancaster University in the UK, this is the first time researchers have examined the trustworthiness of AI-generated faces. And it comes at a time when AI-powered scams are on the rise globally. According to Deloitte, losses due to AI deepfake scams could reach $40 billion by 2027 in the US alone.
The findings of the study, led by psychology PhD student Alexis McGuire from Lancaster, was published in the Journal of Vision this month.
The researchers first asked 169 participants to look at a set of 96 faces — of people of different race, age, and gender — presented randomly, and tell if they were fake or real. The average accuracy turned out to be 58.4 per cent. The fake faces were created using an earlier AI model (GAN) and a newer diffusion model (DM).
Following this, a new group of people were shown the 96 faces randomly and asked to rate them on a scale of one (very untrustworthy) to seven (very trustworthy).
Also read: Not all procrastinators are alike. Cambridge researcher identifies 9 types
The verdict
It turned out that the participants trusted the AI-generated faces more than real human faces. The faces created using the diffusion model appeared to be more trustworthy than GAN faces and real faces. The average rating of real faces was 4.03, while that of DM and GAN faces was 4.70 and 4.36, respectively.
“Our research shows that people are at risk of being fooled by AI-generated images. These AI models have democratised the online space and they are accessible for anyone without technical skills to create fake faces that can be used for a variety of different harms. It is important to inform the public about the ease of creating such images and the potential misuses, and ways in which they might fall victim, for example, through the spread of misinformation, identify fraud, and catfishing,” said McGuire.
“It is critical to understand the threat this democratisation of generative AI brings as well as developing strategies to mitigate potential harms to individuals, organisations, and democracies,” she added.
Tackling online crimes carried out using AI has become a tough challenge for authorities, with some countries introducing new measures to protect people. Japan, for instance, has hired a new AI police chief named AIko to warn people against impostor scams.