Deepfakes, #fakenews and the problem of trust
Posted: Tue Aug 04, 2020 6:16 pm
There's been some drama in the science Twitter world recently when people, who were mourning the death of @Sciencing_Bi from COVID-19, found out that she never really existed and seems to have been made up by BethAnn McLaughlin who founded MeTooSTEM. (https://gizmodo.com/science-twitter-got ... 1844591277 and https://www.chronicle.com/article/did-t ... ill-it-off) This was a long-running con as @Sciencing_Bi was created in 2016, potentially to deflect criticisms of BethAnn's interactions with non-white scientists. Now there's a whole lot to discuss about this debacle but it particularly caught my attention because of something else I read recently…
There was a story about Oliver Taylor, who was particularly interested in Jewish affairs, who accused a man and his wife, who campaign's for Palestinian rights, of being terrorists. The only problem was… Oliver Taylor never existed either! (https://www.reuters.com/article/us-cybe ... SKCN24G15E) Now, I hesitate to call that a deepfake as, while there was some fake profile image, writing under a pseudonym is perfectly normal and accepted practice. The deception, however, is not.
Fake people and imagery do obviously exist. It's easy to make a fake account under a fake name. You can use pictures from all over the internet or make new ones. With increased computing power, you can even make fake videos of real people. That's one thing that is really scary. Anyone can fake chat logs but how are you going to convince people the fake video of you (or more realistically of some political figure) is not actually real? Twitter has verified accounts but recently a 17-year-old Florida boy broke into Twitter and could tweet from verified accounts. (https://www.cnbc.com/2020/07/31/twitter ... r-old.html) We can be picky about our news sources and hope they verify information but, as the Oliver Taylor story shows, pseudonyms, which are vital for some people and not something we should get rid of, as well as attacks that has put fake stories into real news outlets (https://www.fireeye.com/blog/threat-res ... paign.html) make that unreliable as a safeguard.
All of this takes place in the background of a highly-polarised and tribal world where there is little trust and it takes little for people to jump to the worst possible interpretation of anything. Are we moving to a world we can't trust anything or anyone? How are we going to know what is true and not even beyond the standard manipulation techniques?
There was a story about Oliver Taylor, who was particularly interested in Jewish affairs, who accused a man and his wife, who campaign's for Palestinian rights, of being terrorists. The only problem was… Oliver Taylor never existed either! (https://www.reuters.com/article/us-cybe ... SKCN24G15E) Now, I hesitate to call that a deepfake as, while there was some fake profile image, writing under a pseudonym is perfectly normal and accepted practice. The deception, however, is not.
Fake people and imagery do obviously exist. It's easy to make a fake account under a fake name. You can use pictures from all over the internet or make new ones. With increased computing power, you can even make fake videos of real people. That's one thing that is really scary. Anyone can fake chat logs but how are you going to convince people the fake video of you (or more realistically of some political figure) is not actually real? Twitter has verified accounts but recently a 17-year-old Florida boy broke into Twitter and could tweet from verified accounts. (https://www.cnbc.com/2020/07/31/twitter ... r-old.html) We can be picky about our news sources and hope they verify information but, as the Oliver Taylor story shows, pseudonyms, which are vital for some people and not something we should get rid of, as well as attacks that has put fake stories into real news outlets (https://www.fireeye.com/blog/threat-res ... paign.html) make that unreliable as a safeguard.
All of this takes place in the background of a highly-polarised and tribal world where there is little trust and it takes little for people to jump to the worst possible interpretation of anything. Are we moving to a world we can't trust anything or anyone? How are we going to know what is true and not even beyond the standard manipulation techniques?