The Growing Problem of Astronomy & Spaceflight AI Slop on You Tube

[catdlr]

It’s Getting Harder To Know What’s Real

Oct 30, 2025  Neil deGrasse Tyson Explains...

What is real on the internet? Neil deGrasse Tyson breaks down deepfakes, discusses the time a deepfake video tricked Terry Crews, and learns about how AI is used to deceive with the help of Bitdefender Chief Security Strategist, Alex Cosoi.

Timestamps:
00:00 - IT’S FLAT
00:32 - What is “Deep” About a Deepfake?
2:35 - Neil Getting Deepfaked
6:06 - The Stakes of Political Deepfakes
8:11 - Scams to Watch Out For
13:50 - Are We Losing Against Deepfakes?
16:00 - Knowing What’s Real

[Blackstar]

Yeah, surfing various social media sites, I've seen this steady increase--the tech gets better, the disinformation gets easier.

There are weird side effects too, which is that people are increasingly skeptical of real images. I don't think that's unexpected, and in fact it is often part of the point for disinformation and propaganda (erode trust in legitimate sources, and reality). But I've seen it even in obvious situations where somebody posts a video/image that by any measure is legit, but some people still say "It's AI." If they had clicked a few different sources, they would have confirmed that it was legit.



Addendum: just the other day I saw another related example. Somebody had posted space art, and some comments were that it was AI generated. The poster then responded that he was a graphic artist and had created it using various CGI tools. He was polite about it, but I could sense that he was rather mad that he had spent a long time creating something that was dismissed so easily.

[Star One]

Yeah, surfing various social media sites, I've seen this steady increase--the tech gets better, the disinformation gets easier.

There are weird side effects too, which is that people are increasingly skeptical of real images. I don't think that's unexpected, and in fact it is often part of the point for disinformation and propaganda (erode trust in legitimate sources, and reality). But I've seen it even in obvious situations where somebody posts a video/image that by any measure is legit, but some people still say "It's AI." If they had clicked a few different sources, they would have confirmed that it was legit.



Addendum: just the other day I saw another related example. Somebody had posted space art, and some comments were that it was AI generated. The poster then responded that he was a graphic artist and had created it using various CGI tools. He was polite about it, but I could sense that he was rather mad that he had spent a long time creating something that was dismissed so easily.

This is happening more and more. I’ve seen various examples of where merchandise is produced for a property, and because the artist in question has used a particular style in their art people think it’s AI generated. The artist then has to post online their working to show that no it wasn’t AI generated it’s just their style.

[Blackstar]

Here's a good example. He just reposted this from 2023. He spends a lot of time producing incredible animations (not so much still images). And then somebody comes along and gives a prompt to an AI and produces junk.

I don't know what he does for a living, but he has great CGI skills.

[catdlr]

Here's a good example. He just reposted this from 2023. He spends a lot of time producing incredible animations (not so much still images). And then somebody comes along and gives a prompt to an AI and produces junk.

I don't know what he does for a living, but he has great CGI skills.

From my recent perusal around the fighting ring, the hostility between the long-term GCI fans and the recent rise of AI prompters and their "ART" is worse on social media than our SpaceX vs Blue Origin rivalry.

[Blackstar]

I generally despise laziness (except, I guess, when I'm being lazy), and I've seen how AI is contributing to laziness. People go to AI and say "give me a spacecraft image" and plop it in their post about Voyager when they could just as easily go to a search engine and say "Voyager" and get the right photo. Here's an example.

I've recently seen this from a guy on Twitter who should absolutely know better. He regularly posts images of a specific topic (not space), and everything I've seen indicates he knows the subject well. And recently he has been re-posting AI images that in two seconds an expert can tell are wrong. Fortunately, his followers are calling him out on it and telling him that his credibility is suffering.

[catdlr]

I generally despise laziness (except, I guess, when I'm being lazy), and I've seen how AI is contributing to laziness. People go to AI and say "give me a spacecraft image" and plop it in their post about Voyager when they could just as easily go to a search engine and say "Voyager" and get the right photo. Here's an example.

I've recently seen this from a guy on Twitter who should absolutely know better. He regularly posts images of a specific topic (not space), and everything I've seen indicates he knows the subject well. And recently he has been re-posting AI images that in two seconds an expert can tell are wrong. Fortunately, his followers are calling him out on it and telling him that his credibility is suffering.

It has proliferated across all platforms and YouTube content, including "how this is made" videos, military news, space news, etc. It appears that even in the absence of AI within the video content itself, the curator's use of AI-generated 'YouTube Cards' is undermining confidence in their productions. I believe they regard these visuals as 'eye-catching," primarily employing them as a 'clickbait' tactic to attract viewers. However, this approach is having the opposite effect.

[Metalskin]

I generally despise laziness (except, I guess, when I'm being lazy), and I've seen how AI is contributing to laziness. People go to AI and say "give me a spacecraft image" and plop it in their post about Voyager when they could just as easily go to a search engine and say "Voyager" and get the right photo. Here's an example.

I've recently seen this from a guy on Twitter who should absolutely know better. He regularly posts images of a specific topic (not space), and everything I've seen indicates he knows the subject well. And recently he has been re-posting AI images that in two seconds an expert can tell are wrong. Fortunately, his followers are calling him out on it and telling him that his credibility is suffering.

What is crazy, in my mind, is that the Voyager is iconic and how would anyone not recognize that it's the wrong one immediately? Maybe it's a generational thing...

[Star One]

I generally despise laziness (except, I guess, when I'm being lazy), and I've seen how AI is contributing to laziness. People go to AI and say "give me a spacecraft image" and plop it in their post about Voyager when they could just as easily go to a search engine and say "Voyager" and get the right photo. Here's an example.

I've recently seen this from a guy on Twitter who should absolutely know better. He regularly posts images of a specific topic (not space), and everything I've seen indicates he knows the subject well. And recently he has been re-posting AI images that in two seconds an expert can tell are wrong. Fortunately, his followers are calling him out on it and telling him that his credibility is suffering.

What is crazy, in my mind, is that the Voyager is iconic and how would anyone not recognize that it's the wrong one immediately? Maybe it's a generational thing...

I suspect it is generational. It might be iconic to my generation, but I suspect to people born more recently it just isn’t. I can still remember the live coverage we used to get over here in the UK on the BBC children’s show Swap Shop, complete with actual scientists to explain stuff that was being beamed back as well as NASA press conferences being covered. Cannot think we’d get anything like that now. But those were different times.

[Blackstar]

What is crazy, in my mind, is that the Voyager is iconic and how would anyone not recognize that it's the wrong one immediately? Maybe it's a generational thing...

Most people don't know spaceflight or space hardware. I don't know anything about that account, but it may have been generated by a content farm in Asia that simply spews out posts for the clicks to increase advertising revenue. The person involved might have produced dozens of them in an hour, didn't know anything at all about the subject, and may have had ChatGPT creating the text and maybe even the graphics. So no knowledge of the subject matter, and maybe 10 seconds devoted to creating it. Hence, you get junk. I hate that this happens, but that's the world we are in.

There are corollaries to this as well. This image uses the wrong USS Constellation--a scrapped aircraft carrier rather than the still not-built frigate (which is much smaller). However, I've seen people suggest that YouTube channels do this on purpose to get people to write in the comments telling them that they are wrong. This counts as "engagement" and makes them more money. So they're wrong on purpose. That's the world we are in.

[Star One]

Researchers suggested there's more AI generated content appearing on the web than human generated content - Mike Pound from the University of Nottingham talks about why this might be a problem.

[catdlr]

By the
New Yorker Humor

[sanman]

Monsters from the Id



All we have to do now is think it, and AI gives it form

[sanman]

when humans were introduced to AI

[Blackstar]

when humans were introduced to AI

Sadly, yes.

Well, of course it is a lot more complicated than that. AI has value and isn't just garbage. But I'm increasingly concerned about what it is doing that we are not even aware of. I just saw a post indicating that a lot of people (students, academics) are using AI to write papers and it is hallucinating sources such as nonexistent journals. Okay, we knew about that. But what is now happening is that the hallucinated nonexistent journals are now being cited so many times that they are showing up in databases of legitimate sources. The structure of reality is breaking down.

[catdlr]

when humans were introduced to AI

https://youtube.com/watch?v=OxLjvY5CrjQ

Like drugs, once you use it, it's hard to stay away from it.

[Star One]

This can be any kind of history including spaceflight or astronomy.

We used to worry that history was written by the victors. Today, we should be scared for a different reason. For if you want to see the future of our historical record, look no further than the dozens of YouTube and TikTok accounts – many with hundreds of thousands of subscribers and millions of views – which pump out AI-generated “documentaries”.

One popular account I came across does the seemingly impossible by churning out a two-hour “programme” every other day. For any human being or production team – actual historians who care about accuracy, nuance or truth – this pace is physically and intellectually impossible. For an AI, however, it’s just another hour or two (if that) in the office and that is a massive problem.

We are witnessing the industrial-scale pollution of our collective memory. The colourful and intricate tapestry of human history is being shredded and replaced by a cheap, often grossly inaccurate one. And the worst part is that many viewers believe what they see.

https://www.msn.com/en-au/news/techandscience/dystopian-slop-presented-as-fact-is-killing-history/ar-AA1UfTes

[leovinus]

May I offer a different perspective? Instead of all the "the sky is falling and rubbish is everywhere" posts, maybe, just maybe, it is just time to let go of YouTube et al? There was a time when the quality of content was important. No quality, no readers. Today, Google and Facebook care only about the money. As such, any content that make people click on serves them an advertisement and as such generates money for them. Not for the reader/user. Let it go people and find a better platform to discuss and express yourself

We have seen similar issues, on a smaller scale, in the early 90s, with the transition from Usenet to WWW/Html. Html pages and graphics were so sexy and new Usenet was already besieged by spam and rubbish. I remember the neo-hippies posting daily in sci.physics with "I had a dream about the structure of space time and all is clear to me". Really, that was worse that semi accurate AI generated content. Then we had examples like Geocities. Good concept and the whole thing went under with spam, lack of good content, and more sexy alternatives like Fakebook (excuse me "Facebook").

Today, similar stories, YouTube et all go under with the lack of "anything useful". As such, I'd love to see new content servers and sites with factual and accurate stories. NSF maybe? There is plenty of text and video to host on Chris' own site if he chose to do it. I have seen this transition with one or two other sites which I use frequently and it seems to work. There is a subset of readers and contributors who pay and make it work. On you own site, you can moderate.