cliff lampe 03 — Transcript
Full transcript
- 0:00one of the pieces of misinformation we
- 0:01see on social media that has a lot of
- 0:03people worried
- 0:04is the deep fake these are videos that
- 0:06you may have seen
- 0:07where it's a video of a high status
- 0:09politician or somebody like that
- 0:11that's mapped over the face of another
- 0:13person speaking
- 0:14and it looks like the person is saying
- 0:16things they wouldn't normally say
- 0:18now a deep fake is a very heavily
- 0:21computational method
- 0:22by which we map points on a person's
- 0:24face and then we use that map
- 0:26to then take a video from some other
- 0:30source and match the two videos up so
- 0:32what would happen for instance is if i
- 0:34did a good impression for instance of
- 0:35say
- 0:36former president obama i could speak in
- 0:38his voice and you could use the
- 0:40algorithm to match his face on top of
- 0:42mine
- 0:43and make it look like i was speaking in
- 0:45terms of what he
- 0:46and saying things that he wouldn't
- 0:47normally say the beauty of this
- 0:50algorithm is that it pushes the limits
- 0:52of what we understand about artificial
- 0:53intelligence
- 0:54the downsides are pretty obvious right
- 0:56it shows for instance that
- 0:58it's really easy to fake video and you
- 1:00don't have to fake a lot of video
- 1:01if you just get somebody to change a
- 1:03word in the middle of what they're
- 1:04saying it can really create
- 1:06misinformation
- 1:07and harm our strong good information
- 1:10environment that we would like to have
- 1:12another aspect of faked video that i
- 1:14think is important to understand though
- 1:15is the cheap fake as opposed to the deep
- 1:17fake
- 1:18all video is able to be manipulated the
- 1:21cheap fake is where you don't use that
- 1:23high-end
- 1:24artificial intelligence system to fake a
- 1:26video you just use
- 1:27low end common editing tools that almost
- 1:30any high schooler would have on their
- 1:32laptop
- 1:33so here is what you do is you would slow
- 1:35down
- 1:36the rate of speech of somebody
- 1:38especially somebody like me is standing
- 1:40at a podium and talking and is
- 1:42relatively still
- 1:43it would be very easy to slow down the
- 1:45audio without it making it look like
- 1:47my movements had slowed down a lot and
- 1:49make my speech sound slurred and
- 1:50possibly drunk
- 1:52we've seen this a lot in political uh
- 1:54advertising and political kind of
- 1:55attacks
- 1:56where you slow down video by 30 50
- 2:00and the person saying something now
- 2:01sounds like they're either
- 2:03drunk or erratic in some other way
- 2:05another thing of course that's very
- 2:06possible to do is to splice
- 2:08video and cut out key parts of what a
- 2:10person has said
- 2:11depending on where you start a video
- 2:13where you stop a video all these are
- 2:14really strong editorial decisions
- 2:16so taking video out of context
- 2:18reassembling it in different ways
- 2:20so that you have a person for instance
- 2:21who says a statement
- 2:23says a whole bunch of other things then
- 2:25says another statement you remove the
- 2:26middle part
- 2:27put the other two together and it now
- 2:29looks like those two things were
- 2:30associated
- 2:31it's very easy to edit video and it
- 2:33always has been for people who are bad
- 2:35actors
- 2:36to be manipulated the thing that's
- 2:38changed when it comes to cheap fakes and
- 2:39deep fakes is our ability to share them
- 2:41easily across the network
- 2:43because i can take a video that i found
- 2:45and share it with my entire network and
- 2:47then
- 2:47some percentage of my network and share
- 2:48it with their networks it becomes viral
- 2:51very quickly
- 2:51and it's really hard to police and stop
- 2:54these videos when they occur
- 2:55so that's what deep fakes are there
- 2:57they're algorithmically driven
- 2:59video fakes that have existed since the
- 3:01beginning of kind of video media in
- 3:02political narrative they're designed to
- 3:04misdirect they're designed to manipulate
- 3:06and to create an emotional response
- 3:09and they're backed up by these cheap
- 3:11fakes which are really
- 3:12easily manipulated video techniques that
- 3:15almost any high school student would
- 3:16have right now
- 3:17that allow a person to edit video slow
- 3:20it down
- 3:20change it in some way to create a false
- 3:22narrative about what was originally in
- 3:24the video
About this transcript
This page contains the full transcript of cliff lampe 03 by Michigan Online, generated from the public captions YouTube serves with the video. The transcript has 692 words across 120 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.