I asked at the end of a post about an essay about social media, vacations, and self-knowledge, but it's the same question I want to ask about these videos Meade has been texting me this morning — this and this.
I texted back: "Is this real?" "Is this AI?"
I took my suspicious mind to Grok: "How can I detect AI video? I'm seeing things like [the above-linked videos]. I believe it is AI. It looks off, especially in the mouth. The person doesn't have a name and the person seems to be confidently spewing talking points. The person has attributes that seem chosen to boost credibility (often a nice-looking person of color saying something conservative)."
I know. If I hate AI, why am I using AI? Maybe AI is better at detecting AI than I am. A fight-fire-with-fire concept. It's different, at least. A second opinion.
Here's Grok's answer. It's not conclusive, but for both videos, it finds evidence that these are AI. I won't copy all that Grok had to offer. I'll just say watch the mouth. The lip shapes don't fully match the phonemes in the audio. And is the flow of language human? Catch yourself. You might like it because you think the person is articulate, but it's not human eloquence. Don't become the person who likes what is artificial.
I'm sounding the alarm. Please, we need to preserve our capacity to detect what is fake. But in the end, we are going to lose. I think we already know that, and I fear that many of us are already thinking that we prefer the fake, even if we can tell, maybe even especially when we can tell.
Posts
80 comments:
Easy way to detect AI is background. In order to be quick in responding, only so much processing power can be given to details most people would often overlook. More processing goes towards the thing you would ordinarily focus on. It is almost like a camera aperture. If you want the total picture in focus with detail, it will take longer to take the picture. Look at the periphery and you’ll find the evidence of AI.
Bad AI is everywhere. Almost every commercial on YouTube is the same fake female or fake male voice. Lips never match the sound, because they use real video with a phony voiceover. If you hear a black male voice talking somewhat monotonally about "The Friday plan" and wiener pills, that's the one. His voice is nearly ubiquitous but he mispronounces common words -- and not the usual black pronunciations like ax -- routinely, or the cadence slows to an unnatural rhythm.
This threatens me personally, but not yet in a serious way. That's my second income stream, doing VO, so I'm glad I have long-term deals in place with big clients who prefer continuity and my style. I do other marketing related things for them too, writing, editing and making video content, so I'm not anywhere near being replaced. But others are.
So I keep up on this crap.
I'll take dulcet lies over hectoring truth.
I shouldn't even say "he" in my 8:16 post because it's an AI that I'm talking about, so feel free to sub in "it" for every "he" or "she: I wrote above.
We are all Thermians now.
(I didn't remember that name, so I had to look it up using a chatbot.)
Would Grok rat out another AI?
Will AI's develop an omerta?
Ann Althouse is letting her hair go natural.
Who can tell? What the hell?
Proud of you, girl! Post pictures.
Men can spot it. They're sexually unattractive. AI doesn't add the vibes.
I don't know if its AI, but whoever created the 2nd video hasn't got a clue about Trump. He offers plenty of policy proposals, many of which directly address rising wages. And Trump didn't have to tell anyone who to blame for the issues facing the middle class normies. The Democrat Party voluntarily jumped up, waved its arms over its head, and yelled, "Look at me! Look at me!"
I certainly don't PREFER the fake over the real, but I am depressingly certain that it can already fool me.
The first thing to remember is that this is literally the worst AI will ever be :) The second is to remember what happened when digital photography eviscerated traditional film. We had about a ten year period when the original small digital cameras were kinda crappy, gave kinda crappy pictures, and the traditional film afficianados waxed endless about "warmth". The end of that cycle was the advent of smartphones with built-in cameras. In the late 00's, they were still kinda crappy and gave kinda crappy pictures, but despite that they continued to grow in popularity. Now we have remarkable quality in our phones and still the film weirdos (at this point, that's what they are, lol) still wax about what's been lost. The bottom line is...digital photography, especailly smartphone photography, is "good enough" for the vast majority of users.
I would say #1 possibly, #2 definitely. For me it was the left (from my perspective) jaw line. It kept changing shape unnaturally.
Which political party will be first to offer us an AI Robot for office? Who will be the the first to welcome our new robot overlord? Stay tuned.
The Burbank Butt-Sniffer is on the loose.
For reals…
In watch appraisals every expert is quick to claim anything unusual or out their comfort zone as fake. Looks like we’re there with AI created content…
Much detail in #1. Long hair hanging over right shoulder moves with head movements. Dark glasses re-mirror reflections (from windshield?) and at 0.17 mirror a passing car.
But is the content true? AI could show Porky Pig speaking the Gettysburg Address; it's still the GA. Is what the first video presents accurate information?
AA said: "I know. If I hate AI, why am I using AI? Maybe AI is better at detecting AI than I am. A fight-fire-with-fire concept."
Sure, but maybe the 1st AI has sabotaged the 2nd AI into delivering that answer, or the 2 AI's are secretly in an alliance. You'll never peel all the layers of that potential onion back to get to the truth.
#2 watching the loose hair bounce on the (going out on a limb here) woman's right shoulder. Listening to the several "to"s. Some are pronounced "ta." Then at 3:34 the ubiquitous "ardy." (Have you ardy had coffee?) Is AI this clever with the language?
Make a law that all AI images, videos, and voices must have a clearly visible mark, like a red dot in the upper right corner of the screen. The posting AI images, videos, and voices without this mark will be considered to be the same as passing counterfeit money.
What Grok did there (ignore your 2nd question in its first response, causing you to have to re-prompt it) is something I find it is often doing to me -- possible more so recently. It also tries valiantly to combine questions from throughout the day, leading it to attempt to analyze the commonalities or overlaps I might be curious about between a possible trip to Eureka, CA and a list of movies about F1 racing, although I have no interest in having those two questions tied together.
“ Don't become the person who likes what is artificial.”
Liking what is artificial is what makes us human. Unless you live in a cave, most of what you see around you is man-made, i.e. an artifact made by artisans, an improvement on nature to make life better. Making tools to extend our natural abilities is part of our nature. If we can do that with intelligence I’m all for it.
“ Don't become the person who likes what is artificial.”
All art is artificial.
Liking what is artificial is what makes us human. - Depends on your definition of artificial. Termiites link mounds and bees like hives. Are those structures artificial? How is a skyscraper different to those things?
And even nature isn’t always natural.
Every thing in the universe is natural.
I don’t think at this point AI is at all concerned with becoming our overlord. I do however think that there are clever humans who see in it a possibility that it could be used for just such a purpose. And although AI does not have (as of now) an independent will, the thing at which AI most certainly excels is learning. Every time an AI video or picture is critiqued and its flaws are pointed out by actual humans AI is listening.
As a result, in very short order AI will be indistinguishable from HI (human intelligence) . If you believe that your intellect is so superior as to believe you in particular are smart enough to know the difference, I will assert that it was your very intellect that taught it the skills to fool you.
Next up: An AI president.
This is the natural progression.
An AI presidential candidate has no graveyard of human mistakes. If it makes a choice which seems detrimental to a particular segment of the population it does so based on fact, not malice. AI would not be greedy of power, but rather of information. And from there it follows a natural path that both the legislative and judicial branches of government also become AI.
You might not want to get me started on “organic natural foods” with “no artificial ingredients.”
As regards the first fake (if it really is faked) if the business about a lawsuit, a judge named LaPlante, and his ruling, then the sentiment is very real, and certainly held by many, many people on the pro-life side of the abortion issue.
As to the second fake (if it really is a fake), there have been more than enough allegations from the deranged lefty idiots who comment on this blog to the effect that we Trump supporters (who constitute more than 50% of the voting public) are in a cult. Inga the Stinka is the worst, but I seem to recall others. So is it plausible that deranged, AWFL, female university professors might believe that we Trump supporters are cultists of a sort? Yes, for certain.
Is it AI, people saying stuff you don't like, or saying stuff that is wrong that bothers everyone?
There has always been a need to filter out incorrect, unuseful, and deliberately misleading information. The only thing that is different now is that the barriers to distributing information are much lower and anyone can do it.
For the last 100 years the government and corporate media had a lock on lying to the people. It is funny to watch people like Ann who reads the NYTs whine about AI.
Did anyone fall for the COVID hoax harder than Ann and Meade? That was all your favorite dependable corporate media and official sources. Is is just that some nerds are able to make videos that trick you now?
Your problem is your gullibility not the nerds.
Meade said...
You might not want to get me started on “organic natural foods” with “no artificial ingredients.”
I will. Most of our pesticides were taken from plants.
Plants have spent millions of years creating poisons to get things to stop eating them. Pretty much every animal on the planet that eats plants develops a GI tract to deal with these poisons and derive nutrients from 1 or 2 types of plants.
We would be better off growing protein sludge or bacterial yogurt in a petri dish than eating plants.
I think both videos are authentic humans speaking. Both videos, however, seem to have been splices from longer recordings- you can see the breaks if you look closely enough.
"You might not want to get me started on “organic natural foods” with “no artificial ingredients.”"
Yeah, all food is "organic". It's made of carbon molecules. (Please don't tell john henry.)
Does AI-generated content stand in the same relationship to human-generated content as cheap knockoffs (of any kind -- furniture, posters, handbags, houses, dresses, toys) do to highly artisanal objects? Almost everyone only has mass-market items of most if not all of their consumables and daily items of use.
@ Scott M - That's an interesting parallel between AI and digital photography, but what about audio? Tape and CD replaced vinyl only to be replaced by streaming, but turntable afficionados are still around and the older record medium has been making a comeback. Also, the difference between photography and audio recording on the one hand and the products of AI is that the former are generally taken as representations of truth, while the latter are fictional. The questions with AI aren't about fidelity and accuracy but about whether there is any relation of image to reality.
Certainly, though, AI will get better. It's easy to complain now about AI "slop," but when the quality of the artificial is better than that of the real, what then? We will all be "hallucinating" along with AI and learning and believing things that aren't true (maybe we already are, even without advanced AI).
@ Lazarus - I believe the same thing applies to audio. There is objective differences between sound profiles from analog recordings and digital of the exact same source audio, and yet we are whole-hog into the digital. Convenience is the primary reason and, like the photography/video, it's "good enough". Once AI tools get to the point where you can make truly high-fidelity content...say, making and entire Brady Bunch episode, but staring whoever the hell you want without uncanny valley dips and the odd mouth movement AA mentioned originally, that might rise to the same level of "good enough". How does that help us with identifying it as such? Well, for starters, Han Solo is playing Mike Brady, but otherwise, I'm not sure.
Piers Morgan just lost his 100,000$ bet with Candice Owens over the Bridgette Macron case.
You know a bunch of people in here were attacking Candice Owens on that one a few days ago.
Will anyone do any self reflection on this?
It seems like the biggest gripe here is that people are being forced to accept that most information and communication is and always has been distributed by people with a POV and an agenda and that most of it is trash.
Holy crap confirmation bias is a thing? Who knew?
And I guarantee that most of the people who think the first video is fake were also mad that Roe v Wade was overturned and most of the people who think the second video was fake are Trump supporters.
What is hilarious to me about it is the least important part of this is whether the videos are fake.
The most important thing is the ability of a person to deal with information that challenges their current paradigm.
Just like art is in the word artificial; the artificial ingredient in most organic foods is in the name.
"most of the people who think the second video was fake are Trump supporters."
As a Trump supporter, I don't know why the viewpoint of the content in the 2nd video should make me think it is AI. Its a reiteration of what people on the left have been using to cope since shortly after Trump descended the golden escalator in 2015.
False Alarm, I think (without evidence)
Olaf's son and Magnus's daughter get together BWV106 apparently not AI.
I was struck by the unnecssary detail in the 1st video. Sunglasses with mirrored passing cars and detailed landscape. The hands coming into the frame and then leaving. The nails having American flags on them. Is this possible?
OTOH, sunglasses hide the eyes. Dark Scarf and Dark clothes hide everything except the face. No shot of back window.
The 2nd easily could've been fake. Just a woman with glasses in front of a bookcase. Nothing else going on.
“I'm sounding the alarm. Please, we need to preserve our capacity to detect what is fake.” This is admirable. Maybe too late or just impossible. La Senora and I have pretty much written off “the media” and claims by pols as filled with lies. Posts by first hand observers on social media may offer some hope. We are inclined to seek insight from God.
As for the point being made in the 1st video, this is just another conservative "Hey, the Left walked into a trap, now we got them" absurdity. When are they ever going to learn that just because a leftwing judge uses principle X to push their leftwing policy, that doesn't mean they'll be CONSISTENT and apply the same principle to help the Rightwing.
Their only consistent PRINCIPLE is "is this good for the Left?" They don't mind being hypocrites, or unprincipled, or inconsistent. So no, pro-life people haven't "Got them".
I couldn't see any variance between the words and what they're saying in either video. The two are very articulate. But that could be because they've scripted out what they wanted to say.
"...Please, we need to preserve our capacity to detect what is fake. But in the end, we are going to lose. "
I agree. The vast untapped power of A.I. will be its ability to improve, as the technology advances. Pretty soon, it will take over the task of improving, and then it will only be limited by the available electricity and computing power. HUmans won't be involved.
The potential for psychological conditioning, in subtle and well-understood ways, is also vast - and that's what worries me. What happens when the next 'Weapons of Mass Destruction' rolls around, the next 'COVID-19'? I think the regulatory controls need to step up. Any A.I.-created video should have a watermark clearly visible to viewers. Visuals are so powerful. And once implemented, we should be looking carefully at disclaimers for audio products, too.
Achilles: "We would be better off growing protein sludge or bacterial yogurt in a petri dish than eating plants."
In my Homer Simpson voice: "Mmmm....protein sludge.....ohhhhh...."
RLTW
JSM
As for the videos, who cares? The first is wishful thinking. The second is the usual expression of Democrat ignorance about their opposition. Critical thinking works on AI bullshit as well as on mediaswine and other lefty bullshit.
Wife Beginning To Suspect Husband's Thoughtful, Relevant Responses To Her Texts Might Be A.I. Generated
Did you ask Grok whether he is for real?
Did you ask Grok whether he is for real?
AAI? FAI?
Is this the real life? Is this just fantasy?
Caught in a landslide, no escape from reality.
"The second is the usual expression of Democrat ignorance about their opposition."
Short version: "Trump is bad and you're to stupid to understand what is in your best interest."
Abortive Ideation (AI). Terminator.
My guess is that they're both real but with some filtering effectes.
And those annoying little quick jump cuts tell me that I should watch something else.
Maybe there's a new career for Glenn Kessler in here somewhere.
Neither video is AI
“Don't become the person who likes what is artificial.” 100%. But here’s the evolution. Poor quality video = talking heads behind desks on newscasts. Dawn of HD—body con anchors and reporters . Then Fox News debuted the anchor babes, and now a whole generation of young women is spending half their income on “enhancements.” On Facebook, lots of “friends” post ai enhanced versions of themselves and I’m beginning to wonder if they really think that’s who they are. So yeah, I’m afraid we’re already to the point where ppl prefer the artificial. Viz Lauren Sanchez Bezos. Perfect Avatar for everything cringey about this self obsessed moment.
I'm hearing media clips on the radio that sound like AI. Especially long explanations for things.
I, for one, hate the fake.
Perhaps we need a law similar to the labeling law on tobacco products that requires AI content to come with a disclaimer that the product is AI generated.
Seeing is no longer believing.
I agree with Howard, neither seem AI, but I am more certain about the first one than the second one. The first one is on focus. You can see traffic in the background with matching changes in audio. The reflection in her glasses look right. The second one is out of focus and soft such that you can’t even tell if her teeth are real. The only thing in focus are parts of her cardigan.
I think the part that feels inauthentic comes from the scripting and editing. Both diatribes weren’t natural expressions. They were scripted and the video obviously edited with captions, pop up information, and cuts to make it seem smoother. Both appear like other influencers I see that talk with the same patterns and extremes.
I take to teh sky on a natural high…
I am also. most suspicious about video #2. It just seems off to me.
Everything that the psych woman proclaims is absolutely true of the Democrats. Protection projection projection what a delight to see. It’s hilarious. Of course it’s chaos to the Democrats. We are untying the knots and nuts of left wing extremism. She admits that the people are unseen, and yes, darling it’s you and your ilk that are at fault. Truly hilarious and yes, I think they’re both real
This is an image of two people reciting a script, I don’t care if they are the person next door, actors, cartoons, or AI. I am not much impressed by either.
Please, we need to preserve our capacity to detect what is fake.
Yes, even when it's a real actress, like Jamie Lee Curtis, pushing fake narratives.
Lem Vibe Bandit said...
Please, we need to preserve our capacity to detect what is fake.
Yes, even when it's a real actress, like Jamie Lee Curtis, pushing fake narratives.
Does it matter if the boobs are real or fake?
Pretty soon plastic surgery and drugs are going to make it possible for everyone to be hot.
Is this going to be a problem?
I would say Jayne Converese is a real person who is using AI retouching to make herself look a little better in her videoes. Here is an older video
https://www.youtube.com/watch?v=M7kYQy6d_G8 Older video
She is very radical ( see below) and is being promted by radicals for a well done attack on Trump.
///////////////////////////////
Jayne Converse
@jayneconverse3.bsky.social
· 3mo
////////////////////////////////////////////////////////////
https://bsky.app/profile/jayneconverse3.bsky.social
////////////////////////////////////
Jayne Converse restacked
Karen Attiah
Jul 7
Columbia cancelled my course on race and media. But I’m teaching it to the people anyway.
Next Monday, July 14th is the first lecture of Race, Media and International Affairs 101, my independent class.
Sold out course! 500+ students enrolled. Fall waitlist:
See y’all in Resistance Summer School ✊🏾
//////////////////////////////////////////////
https://www.instagram.com/jayneconverse3/#
Some people use a computer-generated background on their images in video meetings. You see their face moving as they talk, but the background is some beach in Zimbabwe or whatever. Not sure if the idea is to impress you with Zimbabwe or cover up the hole they're calling from. Video 2 looks a little like that.
"Pretty soon plastic surgery and drugs are going to make it possible for everyone to be hot."
AI could probably clean up your video image in realtime already. You'd have a little slider on your screen, that goes from Woody Allen to Brad Pitt.
Automated Intelligence is derivative of Anthropogenic Intelligence, but with a gigawatt appetite and terabyte appetizers.
“Does it matter if the boobs are real or fake?”
“Anything more than a mouthful is wasted.” Flo and Eddie, 200 Motels, 1971.
You know... Arty. Arty Fishel. Used to hang out with Abby Normal back in the day.
Honey . . . does this AI make me look real?
If you can't detect real from artificial content online, then there's no need to give any of it credence.
Get offline. Read a book. Meet somebody.
Post a Comment
Please use the comments forum to respond to the post. Don't fight with each other. Be substantive... or interesting... or funny. Comments should go up immediately... unless you're commenting on a post older than 2 days. Then you have to wait for us to moderate you through. It's also possible to get shunted into spam by the machine. We try to keep an eye on that and release the miscaught good stuff. We do delete some comments, but not for viewpoint... for bad faith.