nextMedia.Hamburg presents survey on deepfakes and disinformation

Respondents see deepfakes as threat to media credibility - AI raising many hopes
24 August 2021
Deepfake

Many people in Germany are concerned about automated manipulations of videos, images or sounds, according to the results of a survey by nextMedia.Hamburg presented on Wednesday (August 11, 2021). The results also revealed higher but positive expectations on AI as a technology of the future. This comes against the backdrop of deepfakes, which are trending on social media. A deepfake involves altering a person’s physical appearance using artificial intelligence (AI) and superimposing e.g., their face onto Superman’s body or propelling the victim into Beyoncé’s latest music video.  The fake is typically used maliciously or to spread false information. The authentic-looking video illusions can result unknowingly and simply from tapping the wrong key on a mobile phone app. 

Deepfakes pose risk to credible media

Almost half of Germans or 44.8 per cent fear that they will not (or no longer) be able to recognise fake videos while. 58.8 per cent of respondents are afraid of deepfakes being misused for crimes. And almost three quarters (71.5 per cent) of respondents see deepfakes as a threat to media credibility. What appears to be merely great fun on an app can swiftly become a problem with deliberate deceptions and manipulated news e.g., when identities are stolen or politicians have words put into their mouths that they never said.

More education needed, labelling required

Deepfake algorithms can imitate facial features and voices almost perfectly nowadays. "Until a few years ago, videos and sound recordings were usually considered evidence of reality," said Dr. Nina Klaß, Head of Next Media Hamburg. However, that has changed. "Many of the deepfakes we meet on social media today are developed creatively and for entertainment purposes, but of course there are also negative exceptions to this, which serve mainly to deceive the public and deliberately distort facts," she stressed. For this reason, it is important to educate people about what is already technically possible and to introduce mandatory labelling of deepfake content. Of those who understood the technology, a clear majority (58.1 per cent) were in favour of mandatory deepfake labelling.

Hamburg Next Media AU Survey August 2021
© nextMedia.Hamburg

High hopes on AI 

Regardless of deepfakes, people in Germany are mostly hopeful about AI. Many believe that AI will make their lives easier (39.4 per cent) and safer (25.3 per cent). And around one in three hopes that AI will increase their work efficiency (34.7 per cent) and save time (32.2 per cent) unlike 34.6 per cent who do not have these expectations on AI.

The Civey market and opinion research company surveyed over 2,500 people in Germany between June 28 and July 11, 2021 on behalf of nextMedia.Hamburg. The results are representative of the German population as a whole.
cb/sb/pb