When it Comes to AI, Are We Putting the Cart Before the Horse?
Verance’s announcement of watermarking to fight deepfakes raises a troubling question
With the beginning of the 2024 campaign season and elections just around the corner, video watermarking expert Verance announced in November a collaborative watermarking initiative it’s spearheading along with the help of TV and digital news organizations as well as CE manufacturers to combat deepfakes.
The idea is to give viewers a way to tell if the media they are consuming has been tampered with or modified during distribution.
While on-target—a recent poll showed more than half of Americans are concerned about the impact of deepfakes on the election—the effort raises a troubling question. Does the public even have enough confidence in the media to believe what it reports in the first place, regardless of whether the video is legit or a deepfake?
Consider the findings of a September 2022 Gallup poll. Thirty-eight percent of respondents had no trust in the media. Only 14% of Republicans said they trusted the media, while 27% of independents said they did. Driving the average of all respondents who trust the media to 34% were Democrats, seven out of 10 of whom said they trusted the media. Overall, only 7% of respondents said they had a “great deal” of trust and confidence in the media.
Perhaps Pew Research uncovered the reason why. Polls conducted in February and March 2022 found that 76% of the public said journalists should work to give every side of an issue “equal coverage.” Journalists on the other hand didn’t show a similar level of commitment to “bothsidesism,” as Pew dubbed it. Only 55% of journalists told Pew all sides of an issue should receive equal coverage.
Or maybe it can be chalked up to the public’s general inability to distinguish between fact and opinion. A February-March 2018 Pew Research poll gave respondents five fact-based and five opinion-based statements and asked them to identify what type of statement each was. Only 26% could correctly classify all five statements. Twenty-four percent got four right. The results for correctly categorizing opinion statements were somewhat better, but still poor. Thirty-five percent identified all five opinion statements correctly, and 24% got four correct.
While Verance’s effort to protect the public from deepfakes is commendable, it seems a bit like putting the cart before the horse. It proposes to use technology to solve a problem that is much more deeply rooted.
Before the public can ever trust a watermark the media provides for its content to head off deepfakes, they first must begin trusting the media in greater numbers. Maybe good places to start are more balanced coverage, a hard separation of fact and opinion and clearly identifying each.
Get the TV Tech Newsletter
The professional video industry's #1 source for news, trends and product and tech information. Sign up below.
Phil Kurz is a contributing editor to TV Tech. He has written about TV and video technology for more than 30 years and served as editor of three leading industry magazines. He earned a Bachelor of Journalism and a Master’s Degree in Journalism from the University of Missouri-Columbia School of Journalism.