OPEN AI VIDEO TECH ONE YEAR LATER...

[Ty Harper]

This is what I was getting at whenever I've gone on rants about the changes coming to the way we think about storytelling via 'video' and why our convos around cams are discussions about tech that already has an expiry date: 

 

[IronFilm]

I thought the AI videos we had less than a year ago were mindblowingly good! (even though were also laughably flawed as well) 

So to see such a huge leap in such a short space of time, makes you wonder, what will it be like in 2025? Or 2027?

Stock videography will be dead. (just like stock photography is)

How much more of videography will be dead too, what if anything will be safe a decade from now?

https://openai.com/sora 

 

[Ty Harper]

18 minutes ago, IronFilm said:

Stock videography will be dead. (just like stock photography is)

Yeah he covers this exact topic toward the end, and agrees with you. The entire film and media production industry will be vastly different in the coming years, and that is obviously an understatement. But also the pandemic did not help. I've watched my audio engineer and video peers fight for the tech/quality standards that have governed terrestrial tv and radio since its creation - and the pandemic, which forced all productions to use Zoom video/audio conferencing tech via hosts' living rooms with crappy audio situations - actually proved to the higher ups that audiences don't notice the loss in audio/video production quality. Which is also due to the reality that social media had already crept up and normalized subpar audio/video quality in the minds of the audience. Now YouTube/TikTok/IG audio is a 'style' or 'sound' - not an example of inferior quality. 

[KnightsFan]

A few years ago I said something here on this forum about how AI could even replace wedding photgraphers/videographers. My point was that I didn't know what the technology would look like, but it would eventually be possible. My wild brainstorm was something along the lines of setting up a dozen video cameras, then AI uses that information to generate a whole edit, with closeups and clear audio, nicer lighting, etc.

Doesn't look like the tech that far off. Budget weddings won't even need the video cameras, just a couple photos and a description of what happened. It won't be "real," but will the influencers and influencees care?

[Ty Harper]

8 hours ago, KnightsFan said:

It won't be "real," but will the influencers and influencees care?

Some will - most won't and my kid might wax nostalgic about the days when people took pictures and shot video, but will be just fine with AI, the way we were just fine with mp3s.

[ND64]

I was thinking about gaming industry too.. why I need to hire someone to design water physics, who will probably fail at his job, when AI can generate a realistic AND cinematic image? 

[IronFilm]

2 hours ago, Eric Calabros said:

I was thinking about gaming industry too.. why I need to hire someone to design water physics, who will probably fail at his job, when AI can generate a realistic AND cinematic image? 

AI can't do this in real time, yet. 

So I see it being a bigger threat to the film industry first of all. Doesn't matter if it takes a few minutes, or a few hours, or days, or even weeks, to generate an AI made movie. 

[JulioD]

Other than the cost of labour, how is this THAT different to animation?

it’s not objectively better, it just makes shitty images available to anyone. 

[kye]

18 hours ago, IronFilm said:

How much more of videography will be dead too, what if anything will be safe a decade from now?

I feel completely safe in doing my own home videos of family and friends.

I don't care how photorealistic the AI will get (and it will get to be perfect), there will still be a fundamental difference between what something actually did look like vs what something might have looked like.

This difference will remain as long as people are attached to a physical reality at all.

There are parallels in other mediums as well.  Art forgeries are still forgeries, even if they're perfect.  If you think that no-one will care, just google "art providence" and see how much people really do care.

[Ty Harper]

59 minutes ago, kye said:

I feel completely safe in doing my own home videos of family and friends.

I don't care how photorealistic the AI will get (and it will get to be perfect), there will still be a fundamental difference between what something actually did look like vs what something might have looked like.

This difference will remain as long as people are attached to a physical reality at all.

There are parallels in other mediums as well.  Art forgeries are still forgeries, even if they're perfect.  If you think that no-one will care, just google "art providence" and see how much people really do care.

Agreed. As you say, if people want to use cams and other traditional forms of real life capture for home/family use, no one will stop them. But it's unlikely that media/film production companies in the future will be hiring/paying people who offer camera capture, set design, lighting, etc, etc, as a sole and primary service - which is really what we're talking about. Also, the AI approach won't be seen as a 'forgery' to mass consumers in most circumstances. The ones intended for insidious deep-fake purposes? Yes, of course. But most AI-based video will be seen/consumed as a valid representation of real life ala a painting. It will also be impossible to tell the difference in the future. That's just based on how far a company like Open AI has come in a year. Also, these distinctions we're making around real vs fake will be irrelevant to the vast majority of humans born into it from here on out. All realms of commerce have experienced crushing human labor disruptions in the past and present times (car manufacturing being the most obvious example). What makes this stunning and unique is that it is happening to the realm of commerce (i.e art-based commerce) that we instinctively know humans will continue to do whether they are paid for it or not. You can't say the same for alot of other realms of the human labor economy. So it will be, imo, one of the most poignant blows in the history of human labor to date. 

[Ty Harper]

2 hours ago, JulioD said:

it just makes shitty images available to anyone. 

This is vastly underestimating the quality of AI based video in the near future. You must see that the tech we're seeing/talking about right now will be capable of reproducing imagery that is stunningly life like. The only thing being removed from the equation moving forward is our role in the capture process. But even that is not true - because this AI tech stands on the sum total legacy of everything humans have captured of the world to date. One thing that is humbling about AI-based video/audio etc, is that it is telling us that even our physical existence can be reduced to 1s and 0s.

There were/are human economic systems within which something like AI would/can be used in non-exploitative ways towards human beings. We unfortunately do not exist within one of those system at the moment. 

[kye]

1 hour ago, Ty Harper said:

Agreed. As you say, if people want to use cams and other traditional forms of real life capture for home/family use, no one will stop them. But it's unlikely that media/film production companies in the future will be hiring/paying people who offer camera capture, set design, lighting, etc, etc, as a sole and primary service - which is really what we're talking about. Also, the AI approach won't be seen as a 'forgery' to mass consumers in most circumstances. The ones intended for insidious deep-fake purposes? Yes, of course. But most AI-based video will be seen/consumed as a valid representation of real life ala a painting. It will also be impossible to tell the difference in the future. That's just based on how far a company like Open AI has come in a year. Also, these distinctions we're making around real vs fake will be irrelevant to the vast majority of humans born into it from here on out. All realms of commerce have experienced crushing human labor disruptions in the past and present times (car manufacturing being the most obvious example). What makes this stunning and unique is that it is happening to the realm of commerce (i.e art-based commerce) that we instinctively know humans will continue to do whether they are paid for it or not. You can't say the same for alot of other realms of the human labor economy. So it will be, imo, one of the most poignant blows in the history of human labor to date. 

I agree, but I think there is a distinction here between videos that contain people I know/care-about/etc and people I don't.

If a movie people see has Brad Pitt in it, people probably don't care if it was the real Brad Pitt or an AI version of him, and if they go see a movie they probably don't care if the actors are even real people or AI generated fictional characters.

However, if I watch a video that has anyone I know in it, and it's a depiction of a real-life event then it matters if it was real footage or not.
This might seem to be irrelevant detail, but I think that this means that the following parts of the industry may not be completely gutted:

• Documentaries
• Sports videography
• Engagement/Wedding videography (although some might want a more 'enhanced' version than reality)
• Event videography (birthdays, bar/bat-mitzvah and other religious occasions, etc)
• Corporate videos
• All live-streamed event videography
• News and current affairs TV
• perhaps others?

These are pretty significant percentages of the entire professional moving images industry.  It's easy to start thinking that no-one will pick up a camera professionally any more, but that's just not likely to be the case.  

Even if you're right that people born from now onwards don't have any special relationship with reality (which I don't think will happen for a very long time), the people who are 10 years old now might live for another 100 years and they probably want to continue to want to see real life content, so that will be phased out pretty slowly.

[Ty Harper]

6 hours ago, kye said:

I agree, but I think there is a distinction here between videos that contain people I know/care-about/etc and people I don't.

If a movie people see has Brad Pitt in it, people probably don't care if it was the real Brad Pitt or an AI version of him, and if they go see a movie they probably don't care if the actors are even real people or AI generated fictional characters.

However, if I watch a video that has anyone I know in it, and it's a depiction of a real-life event then it matters if it was real footage or not.
This might seem to be irrelevant detail, but I think that this means that the following parts of the industry may not be completely gutted:

• Documentaries
• Sports videography
• Engagement/Wedding videography (although some might want a more 'enhanced' version than reality)
• Event videography (birthdays, bar/bat-mitzvah and other religious occasions, etc)
• Corporate videos
• All live-streamed event videography
• News and current affairs TV
• perhaps others?

These are pretty significant percentages of the entire professional moving images industry.  It's easy to start thinking that no-one will pick up a camera professionally any more, but that's just not likely to be the case.  

Even if you're right that people born from now onwards don't have any special relationship with reality (which I don't think will happen for a very long time), the people who are 10 years old now might live for another 100 years and they probably want to continue to want to see real life content, so that will be phased out pretty slowly.

I don't think any style or medium or approach is ever rendered completely obsolete. I think the ones you can't monetize just become much less sensible to use for that particular purpose. Whether we're talking reel to reel tape for audio production (which I started with!), OG film cameras, vinyl, cassettes, etc. These become super niche things that creators pull out to return to a "warmer", "more analog" feel. So one day using digital cameras and capturing actual human beings will very likely fall into that category in the media economy - particularly Hollywood.

As far as what you're saying about documentaries. Well that is a world I have been thinking about the most, as a hobbyist and sometimes professional.  I think it will be a mix of what you're saying. For example, I'm doing a family doc about my parents. How they met, fell in love, moved to Canada from the Caribbean, etc. I've already filmed the main interviews with them. But it is clear that there isn't alot of historic video footage of them as kids or as a young couple. So this is totally where AI will transform the possibilities for this project. Because I could take archival photos of them from those periods and use AI to create video representations of those photos in tandem with what they are describing from their past in the video interviews I've already captured. Even create AI video scene of dialogue between them as a young couple based on their accounts. Will it seem strange to them. Maybe? But I'm more inclined to think it will be surreal and fascinating. For me personally, I have no video footage of myself as a kid - and I can tell you I would really appreciate using tech like this for that purpose. 

All that said, I agree that weddings and realtime captures of life and current history will likely still depend on some tyoe of traditional raw video capture. But even in that instance I think there's a scenario where you now use something compact like a phone (with much less interest in thinks like the highest resolution, dynamic range, etc - bcuz AI won't need that kind of stuff) just to give AI the raw data to work with, and then AI reproduces that raw capture perfectly and with even more 'detail' AND gives you infinite angle options to build what we now call a 'multicam' doc of an event. 

I would even think that you could get alot of that done with photos of the event and the people there as well.

If anything audio might be the thing we still need to continue to capture the way we do it now - but again, you prob wouldn't need to worry about having the ultimate quality either - just some raw data for AI to work with.

So yeah, there are ways that traditional tech will likely still be used - but it will prob be vastly minimized and also very obvious that AI tech will be driving the bulk of the creative process. 

 

 

 

 

[Ty Harper]

Also forgot to mention that I think journalism/reporting/news gathering, where the facts of an event are a priority, are spaces where traditional cams/audio, etc will still matter. At least in publicly funded broadcast vs private broadcasting companies. But both realms will be tested by advertisers and waning streams of investment to support those journalistic standards and practices.

[KnightsFan]

10 hours ago, JulioD said:

Other than the cost of labour, how is this THAT different to animation?

it’s not objectively better, it just makes shitty images available to anyone. 

It won't always look shitty. Remember 30 years ago when CGI looked like Legos photographed in stopmotion against a flickery blue screen? Let's wait 30 years on AI generated imagery.

9 hours ago, kye said:

I feel completely safe in doing my own home videos of family and friends.

I don't care how photorealistic the AI will get (and it will get to be perfect), there will still be a fundamental difference between what something actually did look like vs what something might have looked like.

No technology can take away the enjoyment of doing something, though it can take away the economic viability of selling it. Which indirectly affects us, because if fewer cameras are sold, people like you and I will face higher equipment prices.

16 hours ago, Eric Calabros said:

I was thinking about gaming industry too.. why I need to hire someone to design water physics, who will probably fail at his job, when AI can generate a realistic AND cinematic image? 

Certainly AI is already used in the gaming industry to make assets ahead of time. It will be a bit longer before the computational power exists at the end user to fully leverage AI in real time at 60+ fps. When you have a 13 millisecond rendering budget, it's a delicate balance between clever programming and artistically deciding what you can get away with--and that it requires another leap in intelligence levels. Very few humans are able to design top-tier real time renderers. AI will get there, but it's a vastly more complex task than offline image generation.

But yes, AI today already threatens every technical game artist the same way it does the film and animation industries, and will likely be the dominant producer of assets in a couple years. In the near term, humans might still make hero assets, but every rock, tree, and building in the background will be AI. Human writers and voice actors might still voice the main character, but in an RPG with 500 background characters and a million lines of dialog, it is cheaper and higher quality for AI to write and voice generic dialog.

[JulioD]

8 hours ago, Ty Harper said:

This is vastly underestimating the quality of AI based video in the near future. You must see that the tech we're seeing/talking about right now will be capable of reproducing imagery that is stunningly life like. The only thing being removed from the equation moving forward is our role in the capture process. But even that is not true - because this AI tech stands on the sum total legacy of everything humans have captured of the world to date. One thing that is humbling about AI-based video/audio etc, is that it is telling us that even our physical existence can be reduced to 1s and 0s.

There were/are human economic systems within which something like AI would/can be used in non-exploitative ways towards human beings. We unfortunately do not exist within one of those system at the moment. 

I’m going try a different angle.

How is this different to “life like photo real animation that requires no back end labour staring your favourite already famous celebrity skins” ?

and also, last time I checked a Human still “directs” this or arguably “writes” this by prompt, and then refines and edits and selects the best result to “show”

How is this not just a better animation tool?

 

[JulioD]

1 hour ago, Ty Harper said:

Also forgot to mention that I think journalism/reporting/news gathering, where the facts of an event are a priority, are spaces where traditional cams/audio, etc will still matter. At least in publicly funded broadcast vs private broadcasting companies. But both realms will be tested by advertisers and waning streams of investment to support those journalistic standards and practices.

Documentary.  Observational documentary.  Reality TV.

[KnightsFan]

1 hour ago, JulioD said:

How is this different to “life like photo real animation that requires no back end labour staring your favourite already famous celebrity skins” ?

and also, last time I checked a Human still “directs” this or arguably “writes” this by prompt, and then refines and edits and selects the best result to “show”

How is this not just a better animation tool?

I don't think a photo real animation with no back end labour can be described as just a better animation tool. Current animation tools, critically, take years of practice and hundreds of paid hours to create each individual work. A production going from "writer, director, and 10,000 hours of professional, lifelong technical artists" to "writer, director, and a 2 month subscription to OpenAI" is, in my opinion, something to pay attention to and expect disruption from, whether you categorize it as a "just a better tool" or not.

 

Switching perspectives a little, these tools are absolutely perfect for hobbyists like me. I'm never going to hire artists, so my productions go from crap CGI to amazing CGI, and no one loses a job. There are no downsides! If that's the angle you're coming from, then I agree with you.

However, for anyone making a living off of video work, there's a very very large chance that the amount of money that anyone is willing to pay for ANY kind of creative content creation is going to decrease, fast.

[Ty Harper]

1 hour ago, JulioD said:

Reality TV

Reality TV is not even 'reality' and clearly cares little for the tenets of JSP, so I would argue that format will embrace AI with the quickness and you'll have AI-based characters that viewers are rooting for...

[Ty Harper]

2 hours ago, KnightsFan said:

13 hours ago, JulioD said:

 

It won't always look shitty.

Sorry but anyone who's saying this looks sh*tty must not have seen the original Will Smith spaghetti one they had a year ago. THAT was sh*tty - this new one is no where near perfect, but sh*tty is just a ridiculous descriptor of something that is still developing. It is imo good enough for us to understand the implications for the art-based labor industry.