Psychology Today November/December 2020 pp31-31 |SOLUTIONS|”DEEPFAKES Seeing Things. Altered media called deepfakes are upending truth itself. We can mitigate their malicious power-but it may mean letting go of our deeply held preference for visual evidence." By Carolyn Purnell, Ph.D.
Image edited from spectrum.ieee.org from orignal credited to Facebook
For complete detail refer to the article.
Summary of Article
Ben Franklin said “believe half or what you see and none of what you hear” yet as humans we put plenty of faith in what we see based on our ocularcentrism-evolutionarily “eyesight has always been a critical source of information." Think eyewitness testimony, the printed word and video except there’s something called Deepfake these days. “’Deepfakes’ are manipulated videos or images that, as their name suggests, depict fake or distorted events.” They can be created in a number of ways including by using artificial intelligence (AI). Scary that seemingly normal human faces can be fully created images not representing any single human. The technology is so good that deepfakes are hard to spot. Deepfakes typically are used for many purposes only some of which may be for ill intent. We are all probably too aware that the “technology has a sordid origin” notably nonconsenting celebrity “hijacked faces” attached to “bodies of porn stars” in still and video images. Fakes have infiltrated politics as well. An example of a “shallowfake”, that was debunked, shows “Nancy Pelosi slurring as she spoke.” A shallowfake uses traditional audio and visual editing tools rather than AI.
Having said all this, we do suspect that “eyesight isn’t [always] reliable.” For example, eyewitness testimony has been proved wrong repeatedly. What’s making images today less reliable are the use of “sophisticated technologies…designed specifically to trick our eyes.” If we believe a fake image or don’t believe an image each is a problem. For example “Even in the recent case of George Floyd, where most people unequivocally trusted the brutal footage, some argued that it was ‘staged’.” “Mistrust on such a scale with likely further polarize an already dangerously divided populace.” Such disbelief can essentially keep us within our chosen “bubble” like FOX News or MSNBC.
How to improve?
1. Adopt digital watermarks
2. Develop deepfake detectors
3. Criminalize the distribution of deepfakes that aren’t clearly labeled
4. Teach literacy skills-verify all findings from other reputable sources
5. Check our bias
6. Ask ourselves “In what contexts does ‘seeing is believing’ serve us and when does certainty break down”