-Unprotected voting machines & software
-Bots & strawman accounts on social media & fake outlets
-Easily produced content like memes & “shallowfakes” (videos doctored with traditional editing techniques) niemanlab.org/2019/05/what-d…
It’s imperative we have a basic understanding of how easily we can be targeted.
It’s going to be a whole lot worse in 2020 than 2016.
netflix.com/title/80117542
One way to think about these early days of deepfakes as disinformation media is to assume that threats will come from the very top & very bottom.
That theoretically could be a concern for 2024, but definitely not 2020. Why?
Most of the best deepfakes—ie. @jimrossmeskimen +@Sham00k’s recent video or Jordan Peele’s Obama—rely on voice acting for believable audio.
technologyreview.com/s/613033/this-…
But what if a faked video spreads for days before it’s debunked?!
Hasn’t the Internet proven by now that people don’t use fact-checking sources?!
But right now, beyond being hard to make in general, full deepfakes are nearly impossible to make well.
The last thing a troll wants to be is duped. As soon as it becomes clear that a video was faked, that’ll spell the end of that video in the buzz cycle.
1) Their existence used to cast doubt on real, recorded events (what could happen w Spencer, Trump, etc)
2) High-spend counterevidence ops
3) Audio-only deepfakes (+how they dovetail with the above)
cnn.com/interactive/20…
One might even say it’s eerie how little is being done to rectify the security of our elections.
…One *might* say that that isn’t an accident.
(We’re years away from that point—but not many)
Witch hunt. LameStream media. No collusion/conspiracy. (Reminder: the Mueller Report absolutely asserts both).
With the same assuredness Trump claimed Barack Obama wasn’t born in the US, for instance, he could decry the recording as a Deep State deepfake.
Imagine you’re the type of person who believes in a deep state—this is a far more believable lie than a secret Clinton-backed child trafficking ring in a pizza restaurant.
Guess who’ll have the motive, resources, & community to produce high-end faked video content?
(Particularly those who might happen to be, you know, both.)
Speaking purely hypothetically here, folks.
Deepfakes are generally easy to debunk because they leave open so many opportunities for debunking—every muscle movement in the face, every tic, etc. Both our eyes & our ears require convincing.
This is what concerns me about audio-only deepfakes in 2020.
We’re more likely to gloss over discrepancies in audio because we’re used to shoddy audio recordings.
That being the case, there are very simple verification procedures that humans can use to ensure security. digitalguardian.com/blog/social-en…
That said, I’m planning a bunch more of these threads because I’m concerned about technology & media’s role in politics.
If you feel the same, please share resources, tag friends, etc. My DMs are open.
/END