Sora AI and the Future of Video Production | Creatives Grab Coffee 56

Summary

In this episode, the hosts discuss the introduction of AI in video production and its potential benefits and drawbacks. They explore how AI can streamline processes, lower costs, and enhance creativity by allowing for the inclusion of more B-roll content and the creation of different camera angles. However, they also discuss the potential negative impact on job opportunities, pricing, and competition in the industry. The hosts highlight the importance of AI fatigue and the challenges of replicating real people and moments. They also consider the potential need for legislation to address AI-generated video and its impact on misinformation campaigns. The conversation explores the implications and challenges of AI-generated content in video production. It discusses concerns about user perception, organic reach, and the efforts of social media platforms to identify AI content. Legal implications, ownership, and copyright of AI-generated content are also examined. The conversation delves into the potential for AI to enhance existing footage and the need for real people in video content. It explores the challenges of identifying AI-generated content and the potential shift in budget allocation. The conversation concludes with speculation on the future of AI-generated content and the importance of adaptation and navigation in the changing landscape.

Creatives Grab Coffee is produced by Lapse Productions, a video production company based out of Toronto, Canada. Reach out to them for your video production needs.

Video Podcast: Sora AI and the Future of Video Production | Creatives Grab Coffee 56

Audio Podcast: Sora AI and the Future of Video Production | Creatives Grab Coffee 56

Takeaways

AI in video production can streamline processes, lower costs, and enhance creativity.
The introduction of AI may lead to job losses and increased competition in the industry.
AI fatigue and the challenges of replicating real people and moments are important considerations.
Legislation may be needed to address AI-generated video and its impact on misinformation campaigns. AI-generated content may require disclaimers and metadata to inform users.
Social media platforms are working on identifying AI content.
Legal implications and ownership of AI-generated content need to be considered.
AI can enhance existing footage and streamline production processes.
The future of AI-generated content is still uncertain, and adaptation is key.

Timestamps

00:00 Intro
02:17 Introduction to AI in Video Production
04:22 Benefits of AI Video
05:30 AI’s Impact on B-Roll Content
06:25 AI’s Potential for Creating Different Camera Angles
08:09 Efficiency and Cost-Effectiveness of AI Video
08:39 Potential Job Losses in the Video Production Industry
09:21 Balancing Increased Workload and Lower Budgets
10:43 Enhancing the Planning and Pre-Production Process
11:49 AI’s Impact on Animation Projects
12:39 AI Fatigue and the Uncanny Valley
13:51 AI’s Limited Ability to Replicate Real People and Moments
14:14 Negative Impact on Pricing and Competition
15:36 Potential Legislation for AI-Generated Video
27:19 Introduction to AI-generated content
27:58 Concerns about AI-generated content
28:57 Impact on organic reach and user perception
29:29 Social media platforms’ efforts to identify AI content
30:22 Legal implications and ownership of AI-generated content
31:17 Enhancing existing footage with AI
31:57 Potential legal implications for false advertising
32:24 Ownership and copyright of AI-generated content
33:24 Differentiating between fully AI-generated and partially AI-generated content
34:42 AI-generated content in corporate video production
35:08 The continued need for real people in video content
35:53 Challenges in identifying AI-generated content
36:23 Potential challenges in AI software development
37:19 The potential shift in budget allocation
38:33 The role of creativity and consistency in AI-generated content
39:32 The limitations of current AI software
40:51 The possibility of AI replacing certain production processes
42:17 Creating custom AI presets for clients
43:41 The potential for AI to assist in location scouting
46:29 Speculation on the future of AI-generated content
47:20 Adapting and navigating the changing landscape
48:37 Conclusion and call for audience feedback
48:49 Outro

Transcript

Dario Nouri (00:05)
Welcome to Creatives Grab Coffee, the podcast on the business of video production. Creatives Grab Coffee is hosted by Dario Nuri and Kirill Lazerov from Labs Productions. Our goal is to share knowledge and experiences from video production professionals around the world. Whether you’re a freelancer looking to start your own business or a seasoned business owner aiming to scale your company, this is the show for you. Join us as we develop a community of like -minded creatives looking to learn and help each other grow.

Welcome to the business of video production. Welcome to Creatives Grab Coffee. Before we get started with the show, let’s go over today’s sponsors. Do you have a shoot in Toronto? Do you need crew or a strong production partner to help you with your project? Laps Productions is one of the top production companies in Toronto and your go -to video partner. With our strong creative skills and extensive network, we can help you achieve your goal. Laps Productions is able to offer you production services,

white label services or finder fees for project handoffs. Reach out to us on our website at LAPSproductions .com to learn more. My name is Mehran, welcome to Canada Film Equipment.

We are a boutique rental house based in Toronto. We are here to help you guys out with all production sizes. Feel free to contact us to get a quote if you are a production house and you’re looking for lighting, camera packages, or lighting and group plan packages. You can see our contact information in the link below. We are more than happy to help you guys out. Make sure you follow and subscribe to creativesgrapcoffee .com. Thank you.

Hey, what’s up everybody? I’m Matt. Welcome to Audio Process. We are a boutique audio company doing location sound, sound design, post sound, ADR, Foley. We service equipment. We do all your audio needs here in Toronto. We got you covered. Come on down, audioprocess .ca. Don’t forget to like, follow, subscribe, and all of the other internet things to creativesgrabcoffee .com. They’ll be waiting for you. I’ll be waiting for you. And we’re all gonna have a real good time. And now, let’s begin the show.

Dario (02:17)
welcome to another episode everyone. Today we’re going to be talking about a pretty interesting topic that obviously concerns all of us in the video production space and that’s AI and not just any AI, Sora.

two weeks ago they released the Sora AI which if you guys have been living under a rock and don’t know it’s basically a text to video and it definitely got us thinking about what’s going to happen to us in the future.

Granted, it’s just at the beginning, but it’s definitely going to have some wider implications. So anyways, in today’s episode, we’re going to be going over the benefits and I guess the negatives and then I guess where we see our video production future going.

Kyrill (03:00)
But before we really dive into it, let’s just first go over like what our first impressions were when we first heard the news about it. What was like the first thing that popped in your

Dario, when it came in?

Dario (03:12)
that our job is at risk basically. Not now obviously because it’s still early but there’s a… It’s gonna be an interesting future. You know, first they went after the graphic designers and I kind of just went, eh, they’re not gonna come after me and then they came after me. Or technically it was, first they went after people that, well, chatGPT kind of took care of, I don’t know.

Kyrill (03:14)
Yep, it’s done.

Dario (03:42)
First it went after the… God knows what jobs that’s gonna kill. The writers, then the graphic designers, and now it’s us. We’re next on the chopping block.

Kyrill (03:44)
The writers, essentially the writers. Yeah.

you’re forgetting also the photographers because that was the next bit with the AI photos that came out. And it got me thinking a lot about like kind of like how it’s going to affect specific people within our industry and especially what industries it will affect. But that’s kind of like something a little bit more later on that we can kind of go over. But what would you say are some of the really good aspects of the AI video coming in now?

Dario (04:22)
There’s quite a bit actually. So again, at first impressions was, oof, we’re next. But when you look at it from a business perspective, all these AI tools that are coming out essentially are just that. They’re just tools. So they’re just helping to.

streamline our processes in a way. So and you know what lower costs as well. They’re very cost effective, right? So the way I’m looking at it is that it’s going to help our business in the long term, right? Because again, well, let’s go through like the very first one that I can think of on a technical level is like, okay, now we can include more B -roll content.

in our shoots right we can just like not now obviously because it sort of just came out but maybe in the long term what we could do is just like let’s say we’re out of shoot uh we don’t have time to grab something well we don’t have to stress out we could go home and then just text a video it and then boom we have that b -roll shot we needed for the video there it is that’s one really cool thing we could possibly do in the future

Kyrill (05:28)
Well…

It’s a potential, but again, it also depends on what you’re capturing. And this was like something that we’ve discussed before where if it’s specifically B -roll that might be even like taken from the stock footage world specifically, because a lot of the time when you’re filming content for a lot of corporate clients, there are sometimes footage and clips that you need to source from stock footage sites.

that are very specific to what you’re talking about, something that may not be able to be captured at the location that you’re filming. And this is something that could essentially affect that aspect of the whole process in terms of sourcing B -roll content. I wonder how advanced it’ll get to the point where if you feed it, I’m curious if you feed, for example, some B -roll that you’ve already filmed from a certain shoot and asking it to say, give us a different angle.

with from this scene that we’re filming. It’s like, oh, can you include a wide from this scene based on the footage that we gave you? That I’m really curious.

Dario (06:32)
Well, you can already do that in Photoshop. You can already do that in Photoshop. I was speaking with a wedding photographer and he said that he took this shot of the bride and groom and he’s like, oh, you know what? I wish that I shot a little bit wider. So all he did was he used, I don’t know what the tool is in Photoshop, but you can just like zoom out and it’ll create the missing stuff in there, the missing contents. So we could probably do that at some point with the Sora AI. That’s lovely.

Kyrill (06:36)
I know.

Yeah, well they’re also doing –

They’re doing it now, Dario. Dario, they’re doing it now already. I don’t know if you know, but like a lot of people even take scenes that they filmed in video and then they go into Photoshop to extend that frame. As long as you don’t want your subject to kind of go out of the normal frame, you could actually get it a little bit wider in terms of that. But that’s…

Dario (07:16)
Yeah, I’ve seen those on reels. They’ll do movie scenes and they’ll just extend the top and the bottoms. But they’re not really doing video out of it. They’re doing… It’s just an image. Yeah, once they can do video, that’d be pretty… That’d be a lot more interesting.

Kyrill (07:26)
That’s that’s what I mean. Like.

Yeah, that’s what I was touching upon. More so specifically, like say for example, you filmed a whole sequence of shots where you have like a nice wide, a nice close up or something, but it’s like, oh, can you get a side angle shot of the subject that we’re filming? That would be cool because essentially it’s kind of recreating a different camera angle as well. Imagine even like if you’re doing an interview, you do one camera and then you ask Sora to create a secondary angle with it. If it can get to that point, then,

That is definitely going to save a lot of budgets in future video projects. I’m curious when it’ll get to that point.

Dario (08:09)
Yeah, like I’m thinking if we’re doing like an interview shot, we can just ask it, okay, can you whip up the secondary angle? Or, you know, it’d be really cool if you already have the secondary angle. It’s like, can you add movement to the secondary angle, right? Like just have it be like on a slider, have that kind of movement. That could be really cool. It sucks for, you know, the videographer or cinematographer that’s not gonna get put on that project as a secondary. That’s gonna suck for them, but…

Kyrill (08:20)
That’s what I was saying.

Dario (08:39)
The way things are going, like with content, with the cause of projects, like the budget’s dropping and people just wanting faster content outputs anyways, like that’s probably the future it’s heading in anyways. So.

Kyrill (08:53)
Well, the balancing act is also where there might be less people being hired for certain jobs, but there’s probably gonna be more jobs as well because exactly what you touched upon, there’s always gonna be a need for more content. And the more efficient projects can become, the more efficient you can be with your budgets, the more potential there is to actually create projects. Because you were saying one of the…

Dario (09:18)
That’s not good.

Kyrill (09:21)
Yeah, I know that, but I’m just saying that is what is kind of like headed, but like if there’s more work going around, people will still find work, but that is one potential. Let me get to the next point.

Dario (09:29)
Yeah, but, but, but there’s a lot of, if it’s low budget stuff, it’s not, it’s not gonna, like if the, if it lowers the barriers to entry. Yeah, yeah, yeah, yeah, yeah.

Kyrill (09:38)
I said efficient, I wasn’t saying lower. I wasn’t saying it has to get low. I was saying bigger projects can become more efficient and you can get more of those bigger projects. Yeah, it’s slightly reduced budgets because what we were talking about originally was also how it streamlines our production process and the reason that budgets are gonna become, the reason people are gonna be able to save on budgets and pricing is because.

you’re gonna be saving time essentially as well. That is one of the biggest thing. When there’s more time being saved, then it’s also justifiable to be able to, for a lot of us on the creative side, to be able to lower the budgets. It’s like, okay, if people are trying to make their, if budgets are gonna go down for projects, the time aspect has to also be, how do you say?

has to reflect that as well because you can’t be like lowering budgets and then increasing time in terms of work, right? Because then no one can do it at that point, right?

Dario (10:43)
We’ll see, because it depends on how easy it is to prompt the software, because if the prompts take a while, then… Yeah. So basically cost -effectiveness, like you can lower your project cost, that’s one benefit. Enhancing the creativity, like, and by creativity I mean really just being able to capture more of what you wanted to capture, not having that time constraint in the first place, because obviously when you’re doing production, you’re there for how many hours you are.

Kyrill (10:48)
That’s my point. That’s my point.

Dario (11:13)
Usually it’s like 10 to 12 hour day and it’s hard to fit everything in at the same time and then I Guess that would be it. I wouldn’t really I don’t really see more

Kyrill (11:20)
Well, the other, no, no, no, there’s more. There’s a lot more. There is more Dario. Like one of the other aspects is in the planning stage and the pre -production process because sometimes when you’re coming up with concepts with a client, you need to come up with some creative ideas and even create visuals, mood boards, storyboards. Those things take quite a bit of time. And especially when you’re like right now, like when you want to create a storyboard based on reference images, you have to look for those very specific reference images.

There are tools out there like FrameSet where you can Google specific frames that can kind of help you with that. But again, you’re pulling from past projects, past films and things like that. What this could potentially do is create a very unique storyboard that’s based on a client’s image, a client’s message. Like I’m curious if it gets to the point where you feed it all the different like branding guidelines and.

visual elements and imagery from a client that you can kind of generate a unique storyboard that’s tailored to that specific project. There’s a lot of potential in the planning stage for it, which I think is great. I mean, this also kind of goes back to, you know, AI photo that kind of can help with that, but the AI video aspect is like you could actually see it in motion. It’s like, how does this feel? You can pair it with.

with some potential music options. It’s like, see if we filmed it like this with this song, this is the kind of tone and energy you can kind of feel. This is before you even get to the video filming part of the process or especially for animation projects, it’s going to really be revolutionizing things because that is where I think will be the first industry that gets hit really hard with it before it gets to any other ones because that’s the one where a lot of digital imagery is created from scratch rather than.

being filmed, right?

Dario (13:12)
Yeah, so yeah, I guess the pre -production aspect as well. That’s for sure. Maybe, yeah, I guess all aspects are going to be affected. It’s going to be interesting. Interesting. But let’s move on to the negatives now.

Kyrill (13:26)
Oh, before we go to the negatives, there was one that we forgot to touch upon was that the one kind of silver lining in the future is that that will kind of keep a lot of jobs is the fact that it’s going to be hard to replicate real people and real moments. So I feel like a lot of events capture is still going to be very much in full swing because you’re trying to capture real people. And.

I think that’s one of the industries that’s not going to be affected as much. I mean, it can maybe enhance in certain situations, but there’s still going to be a need for people to capture events as they are happening. And so that is one silver lining is that that industry will still require people in our space will still need jobs. So, you know.

Dario (14:14)
Okay, let’s move on to the negatives, which there are many. I guess we’ll start off with the fact that, I mean, you mentioned how if videos become more cost effective and, you know, like the budgets will get smaller and then you said that there’ll be more work to go around because it’ll lower the barriers to content creation for clients. I actually think that’s a very big negative because if it’s cheap to create content,

Kyrill (14:42)
Mm -hmm.

Dario (14:44)
then you’re just going to work at bulk and it’s hard to monetize bulk especially in service. It’s very hard. So that’s going to be a major, major problem because if anyone can just pick up a computer and a camera and just quickly create a video for someone and have an AI kind of guide it through the whole story and then, you know, have another AI creating the story visually for them.

I mean, how can we justify our pricing, right? So as an industry, that’s going to be, it’s going to be tough. It’s going to be tough to navigate, to justify telling a client, you need to pay me this because this is what I output. Especially when you consider the fact that a lot of clients can’t even, clients and leads, they can’t really tell the difference between good video and bad video. So we’re already struggling on that side. Now you mix this into it and oh my God, now it’s going to make the sales pitch a lot tougher than it already is.

Kyrill (15:36)
Yeah.

I think sometimes it’s not so much the fact that they can’t tell the difference between good and bad. I mean, that is the case a lot of the time, but a lot of the time it’s also just a matter of this is our budget, what can we get? Oh, we can get something great, that’s all we need. Sometimes some clients, they just need to get some form of video content, they don’t need it to be Spielberg quality. And by that I mean, like for example, if it’s a difference between like say 90 %…

and like 70 % in terms of quality, if the 70 % is significantly cheaper, they’re gonna always go with that number because that’s what they can afford. It’s like, if you start looking at things from a business perspective, you can understand why they go for certain decisions a lot of the time. It sucks, but.

Dario (16:30)
Well, no, I don’t understand it because, again, why spend the money, get a bad product, and then not market it well? It’s a poor reflection on the employee that’s doing it, to be honest with you. Because a lot of times, yeah, they just don’t know any better. But I feel like a lot of times they might not even care because, again, it’s not their money. If it was their money, I would be a bit more diligent than…

Kyrill (16:41)
No, that I don’t get, but it’s like, you could understand, you could understand their reasoning, but it’s not good.

That’s my point. Like they don’t care.

Dario (17:00)
than I see. But yeah, it’s definitely going to be a big problem in our industry. I mean, it’s already an issue now because you have the low barriers to entry in terms of gear alone means that we’re dealing with more competition. I mean, granted, we did benefit from that. But again, if it lowers, if it increases competition, it means the marketplace is going to be tougher.

Right? For you to sell your services. Like sometimes you see it as well. It’s like, oh, well, these guys are charging way less. And it’s like, oh my God, now we got to compete with that and tell them why we’re better. It’s just, it’s just a hassle. You know?

Kyrill (17:26)
Yeah. To be honest.

Yeah, I mean, various are entry are always gonna be getting, they’re always gonna be getting easier. And they’re gonna always keep lowering and it’s gonna be easier and easier to get into the industry. Like when we got into video production, back then it was considered very easy to get into. And it’s like, now you look back versus like how easy it is for people now to get into it compared to back then. It’s like, wow, like we had a lot of challenges that we had to circumnavigate compared to what it would be like now. And then…

now with this big leap in AI video, it’s gonna have to force everyone to adapt completely. Like AI, I mean, video animators especially, they’re gonna have to completely pivot their service offerings and their skillsets. They’re essentially gonna become AI video prompters, right? They’re gonna have to really develop that skill. It might actually help make their processes a little bit more efficient as well in terms of creating their content in bulk maybe, but.

Like that’s one of the biggest shifters, because I think, as I mentioned before, that is the one industry that one part of the industry that’s going to be affected first and probably the hardest. Um, but I mean, there’s a lot of other potential negatives as well. And like, uh, we already see this with a lot of other aspects in, in, in the world is AI fatigue, right? When you start seeing more and more.

AI video content, I feel like you could always tell, kind of like how you can tell in CGI movies that it’s CGI. It’s like, yeah, it looks really realistic, but at the same time, you can tell. You can tell.

Dario (19:08)
It looks good. Yeah, it looks good at the beginning and then after like even with Sora it looked really good the first day it came out. I was like, ooh, it’s looking really good. But then after like a day or two, you’re like, yeah, you can kind of tell.

Kyrill (19:20)
When you keep looking at it, you notice things. Like that’s the thing. Like I think, I mean, again, it’s in its early stages, right? It’s probably gonna get more advanced, but at the same time, I feel like we’re still gonna be able to tell. We’re gonna be able to tell when it’s AI, you know, based on maybe like what the people are doing. I hope, I hope. I think so, but I’d like to think so.

Dario (19:38)
I hope that’s the case, like, cause uh…

Yeah, I hope that’s the case because if not then yeah, that’s that’s gonna be a tough one I mean again, we’ll be able to create some really cool content with it if it gets to the point where it passes that uncanny valley but Again, it will affect this negatively as well and you know, we already mentioned how We’d be bringing less people to to to set as crew members. So there’s already going to be a negative impact to

to industry professionals in that sense as well. Like I’m thinking not just also from like Sora, but if you think about, for example, audio, you know, the audio tools out there right now, those, the AI tools for audio, they’re pretty damn good. So sometimes like if you’re, if you’re trying to close a deal and you’re like, okay, the audio ops costing me, it’s bringing it over budget. And I just.

Kyrill (20:26)
Mm -hmm.

Yeah.

Dario (20:40)
If I can just cut that out, I can close the deal. You’re gonna cut that guy out and you’re just gonna go like, okay, I’ll just use regular audio that I have, like a Tascam or whatever, a TGI mic, and then I’ll just clean it up and post. And if it does like a pretty good job, well, a lot of people are gonna make that decision. So that’s one crew member that’s gonna have a tough time. We already mentioned the secondary cinematographer, second cam off is gonna be tough. Like, it’s gonna get scared.

Kyrill (21:01)
And that’s.

I mean, if you look at a lot of like historically in terms of jobs that would have to be cut out in terms to fit within a certain budget range for a project, like if you look back in the day when they were creating video content, you would need quite a few people, even today for a camera department, right? Like you would need a focus puller, you need another camera assist or like first, second or third camera assists. But there has been leaps in technology, like for example, like for…

lower budget projects, you know, a lot of corporate based projects sometimes where even just autofocus alone, that kills one entire job right there because you need to make it work. And Sora’s introduction just kind of reminds us how every leap in technology is going to cut out a job somewhere. But at the same time, I feel like it could potentially create new jobs, right? Because essentially what people, what we’re gonna have to become,

Dario (21:58)
Mm -hmm.

Kyrill (22:06)
is AI video prompters. That’s gonna become a new job essentially, right? And I mean, I don’t know, like we’ll see.

Dario (22:12)
So if you got into this job because you failed at essay writing, well, sorry to tell you. You gotta get back into it.

Kyrill (22:19)
You gotta get better at writing again. Or it’s a give me video.

Dario (22:23)
Or you just tell ChatGPT to write the prompt for me please. I’m illiterate.

Kyrill (22:28)
It actually has that already. It actually has that already. When Sora came out, or when Sora was announced, I looked up what was already kind of out there, and Chad’s GBT already has a Sora prompt aid. So you give it a little bit of a general description of what you want, and then you ask it to create a structured prompt for you, and it gives it very well detailed. It’s like the tone, energy, feel.

Dario (22:54)
You know, you mentioned the prompts, like we have to become prompt engineers. You know, actually they’re working towards going, I forget what the term is for it, but, cause you know, prompts have to be written in a certain way. Like I’m remembering like my mid journey prompts, like you can’t write like a whole paragraph. You have to write a couple of words, comma, like this, that, that, that, that, that, like that has to be broken, right?

Kyrill (23:24)
No!

Dario (23:24)
Kind of like bulleted in a way. So I know they’re trying to shift away from that to like more free flow. So you could just say something as simple as you just describe it in a paragraph type of situation. Right. So that is the next. That is what they’re trying to do. I think, I think stable diffusion is trying that, which means probably mid journey will introduce a mid journey introduces it, then probably it’ll be included in Sora. So.

Kyrill (23:26)
Bullet point, yep.

Mm -hmm.

Dario (23:54)
Majority prompt, I mean prompt engineers. Maybe, who knows? It’s evolving so quickly. Again, just think about it. Last year they were doing the Will Smith feeding spaghetti video. That horror film that they did. And then now a year later they’re already at this stage. So God knows in a couple years where it’s going to be at. So maybe even the prompting might be a lot easier. Which again, reduces barriers to entry. But there’s another point I want to bring up, which I think is going to be very important.

I think a lot of people forgot about this. Kero, what’s so special about this year?

Kyrill (24:31)
Oh yeah, the election year. Yep.

Dario (24:33)
Okay, so interestingly enough, they released something that can be a boon for nefarious actors that want to introduce misinformation campaigns, right? So as soon as I saw that, and I remember there was an election year, the first thing I thought of was they’re 100 % gonna try to include some type of legislation for AI generated video, right? So as soon as that happens, I…

Like it’s probably going to say something like, oh, you got to, you got to have a watermark on the video saying it’s AI generated or a disclaimer at the beginning or end. Right. So imagine if you’re creating like a corporate video and like they probably, they have to follow regulations. So they’d have to include a disclaimer for AI video at the beginning of the videos. So as soon as I thought about that from the legal side, I’m like, they’re pro like most corporate clients. I mean, I think the, you know, the big ones, they’re probably.

Kyrill (25:10)
But how would they do that? How would they?

Yeah.

Dario (25:29)
going to have something against using AI generated video in their video deliverables. Cause I know some of our clients, they can’t even use chat GPT at work. They’re not allowed to use it because you know, it’s information shared with the system and all other stuff is proprietary, right? So probably would AI video be the same thing.

Kyrill (25:32)
Yeah.

This.

Yeah.

it’s gonna be like those disclaimers like for infomercials in the States, like it’s gonna first be like, this video may contain AI generated video content. And it’ll just basically be almost like the pharmaceutical industry, how they advertise in the States where it’s like just very generic, happy footage, but like then listing out all the…

Dario (26:01)
Yeah, yeah, yeah.

Kyrill (26:15)
all the side effects that a lot of people might get from like buying that drug, it could have a very similar stigma where it’s like, okay, if there is like, this video might have AI generated content. It’s like, okay, what is now what’s my opinion gonna be going into this? It’s like, am I gonna immediately start assuming that everything is AI generated? Because then if like, because like you said, because of this,

Dario (26:20)
Yeah.

Yeah.

Kyrill (26:41)
the news is gonna become even more untrustworthy because you’re not gonna know what’s real or fake anymore. And if you have to include a watermark on AI video, you can’t really crop it out because I’ve seen some sources like ColdFusion, they touched upon like, oh, what if we just put a watermark on the video? But it’s like, yeah, but then you could just crop that out, you know, if it’s in the corner or something like that. For a watermark to work, it would have to be like dead center.

just completely covering the whole image and then it’s like, yeah, then what’s the point?

Dario (27:10)
No, I think they’ll, I don’t think they’ll do it like that. I think it’ll just be something at the beginning. Like, uh…

Kyrill (27:14)
Yeah. Yeah, that’s a disclaimer. It’s not a watermark. It’s the difference, right?

Dario (27:19)
Yeah, I don’t think they’re gonna do what or maybe they could put something in the metadata Maybe I don’t know but it’ll it’ll 100 be something at the beginning of the video saying Following is AI generated content. It’ll have to be like that. They’re 100 % gonna do something like that

Kyrill (27:25)
But…

Yeah, doing the metadata, I don’t think is a trustworthy way because they had it as an example with AI photo content where if they put it in the metadata of photos, but then people just screenshot it, that eliminates all that metadata right there. So then it’s very easy to circumnavigate metadata, I think, in that type of situation. But.

Dario (27:58)
I think it’s not gonna be a watermark. I think it’s gonna be, because a watermark would be too distracting. Or unless it’s something hidden in the image that you can easily tell. But I think it’s gonna be like the drug commercials. I’ll put like a disclaimer at some point. Like again, this is for like stuff, this is like B2C type of stuff or public corporate content. Let’s see how it goes. Another thing is,

Kyrill (28:01)
Just disclaimers.

Yeah.

Dario (28:28)
If you’re publishing AI content on social media platforms, I know right now on TikTok, you got to tell it it’s AI generated. So, and I know AI generated content doesn’t do as well organically. From what I’ve read, I don’t know if that’s accurate or not. Again, I’m not the best marketer, but that’s definitely gonna impact the client’s decision to include AI content in their videos. Because if it’s not gonna do well organically or even with some, like, you know, whatever, like,

Kyrill (28:44)
Well, what –

So what TikTok? So what TikTok?

Dario (28:57)
they might be more hesitant to use it.

Kyrill (28:59)
So what TikTok is doing right now is they’re trying to train their own AI identifier, where essentially it’s asking users to tell it when it’s uploading content, what is AI content and what isn’t so that its own AI detection system can learn how to identify AI content. So it’s right now gathering data from that aspect and that is how they’re gonna train these AI detection systems where people tell what would be AI.

generated and what isn’t, you know, we’ve mentioned how like some, sometimes people make comments like, Oh, this has to be AI where it may not be the case. It would have to be up to the uploaders to identify it more accurately. So I think it’s going to get into the, into the, uh, I don’t see it as an impossibility that a lot of social media platforms are going to be able to identify what is and isn’t, um, AI. It’s kind of like how you see what, what.

people identify as misinformation these days online. It’s like, oh, this video talks about this topic. So it may contain misinformation or this and that, you know, like there are forms of that already in regards to certain content. So this is just another medium essentially that they’re gonna have to try to identify and sort through, but I don’t know. We’re just a couple of weeks in.

Dario (30:22)
But anyways, from the corporate side, there, again, I can see clients being hesitant to use it if it’s not going to perform as well as a non -AI generated video. But the interesting thing would be if it is fully AI versus partly created by AI. That’s going to be the distinguishing factor. Because then again, as well, like what if it’s like 99 % real, but you just used…

Kyrill (30:33)
Exactly.

Dario (30:52)
like a little bit of AI, like you were missing a shot and you just put it in. Would you have to declare it as like a fully, like, you know, like this is an AI generated video. That’s, that’s the interesting part though. I’ll have to have like different tiers. It’s like, this is partly generated, this is fully generated. And then again, like the thing is it’s a problem if these social media giants decide they don’t want to push it, which.

Kyrill (31:17)
Exactly.

Dario (31:18)
could be the case, but it could also get to the point where it’s so common they might decide not to punish users for generating AI content. Could also be that, who knows?

Kyrill (31:28)
It could also enhance. I’m curious if it’ll get to the point where you feed it a video shot and you ask the AI software, enhance this to look clean and polished and things like that, almost like the post -production side of things. Do a little bit of color correction and things like that. I wonder if it will have that capability as well where, to your point, because AI was involved in creating that aspect,

will it be flagged as AI content at that point, when it comes to enhancing the content. But I mean, one of the other potential legal implications is if you create an AI B -roll or content for some kind of organization and that kind of like promises something to the consumer and then the consumer purchases a product or service based on that false.

Dario (32:01)
Who knows?

Kyrill (32:24)
selling. I don’t know if that will potentially be something. It’s not ours, but it could, but for corporations, it could be.

Dario (32:25)
That’s not our problem to be honest with you, as a video production. It’s not our… It’s not… I mean, to be honest with you, every content we create goes through so many levels. Whoever screws that up, it’s up to them. That’s not really a video production. That shouldn’t be in our…

Kyrill (32:42)
No, it’s not for us. It’s more so on the client side, right? Like it’s, it is a, it could be a potential thing. I know I’m just saying that that is a potential thing from their end that they’re gonna potentially be worried about, right?

Dario (32:46)
That’s their problem, to be honest with you. That’s their problem.

Yeah. I’m just saying it has nothing to do with us. It’s one of those things that’s like, who cares? It’s not our problem. Ownership. That would be that would be something that could potentially be a problem because who would own the content, right? If you fully generated on like a platform like Sora, it’s not owned by you or technically, I think how it works. No. Yeah. So if you because I know for mid journey, like you don’t own it or you have ownership, but.

So does open AI, they also own everything. Yeah, you have like a license to use it or something, or they say you own it. I don’t know. It’s like a little tricky in that regard. But it’s that I can see clients being wary just because they won’t be able to own the content. If it’s B -roll, I don’t see it being an issue, but like, let’s say you need to create. I don’t know, like imagine you got to create like a like a person.

Kyrill (33:26)
It has like the full license. Yeah.

Mm -hmm.

Yeah.

if it’s general.

Dario (33:51)
And you know, you’re already using AI voiceover and you’re just matching the two together. That could be a problem because then they don’t own the content. So their legal department might tell them they’re not allowed to do that. Because again, oh, and also because AI generated content on these platforms often is you don’t own the copyright on it. So other people can just take it from you. Right. But I’m trying to think like from the corporate side, like who knows?

Honestly, I think to be honest here, I think we’ll just be using it on the corporate side because we do a lot of B2B B2C Not so much B2C. I think it’s more B2B that we do. I just see it enhancing whatever we already have That’s how I really see it. I don’t see it like getting to that point where it’s like, oh my god these bigger philosophical and legal questions, right?

Kyrill (34:42)
Yeah, I mean, like one actually benefit that we didn’t touch upon is that the AI video is not going to replace the need to capture, you know, important people and organizations. A lot of the time people need to hear real messages from the real leaders of like a, like the CEOs, the CMOs of organizations where they’re addressing the public, other clients, their employees.

you still need to capture those people, you still need to film them. So it’s not going to replace the need for that type of content either. And yeah, it could probably help enhance the process for capturing that content. It can help give us better B -roll stock footage and things like that if it’s needed for a certain project, especially cutaway shots, transition shots and things like that. It can really help in that sense of the process, but the real meat of the content is still.

If you have to capture real people, it’s still gonna be a need for us there. So that’s the one silver lining going forward is that people are always gonna need to hear things from real people. And if it’s AI generated, you’re gonna be able to tell, I think, so.

Dario (35:53)
But you know what the caveat is you’ll only be able to tell if you know exactly who those people are

Kyrill (36:00)
Well, yeah. But a lot of the time, these are like, especially internal videos, you’re addressing very specific people, especially when you’re creating content for a corporation to send out to their clients. They’re going to know who they’re hearing from. So I’m sure people will understand what’s real and what isn’t. I mean, again, who knows? I’m trying to be optimistic here. Like it could get to the point where, but that could be a problem is like they create, if you create AI video,

Dario (36:01)
Ha ha ha.

You know what?

Kyrill (36:29)
of a CEO saying something that he didn’t could maybe potentially they might say something and it wasn’t approved by the CEO. What happens then? You know, like.

Dario (36:38)
Carol, that’s not gonna be an issue because stuff, especially stuff relating to the CO, goes through so many layers.

Kyrill (36:44)
That’s yeah, that’s that’s the one thing though.

Dario (36:46)
There’s so many people writing like saying this is okay. You know what I was thinking would be really cool is like…

Yeah, you’re there filming the CEO or whatever and like you just wanted to add like an extra light that day, but you didn’t have it wait, or you just didn’t have, you didn’t have it wait year because of the environment you couldn’t fit it in. You could just feed it into like, so or whatever and say like, add a hair light on the left side, matching the intensity of the one on the right side and it does it. That would be so cool.

Kyrill (37:19)
Yeah, well that goes back to what I was saying before where like the, I wonder if it’ll get to that point where it enhances the current footage that you feed it. It’s like, here is the shot. Like it should, it should get to that point. And that will help in the post -production side of things sometimes, especially if when I feel like it’ll really come in clutch, this AI software when you need to do fast turnarounds, especially. And the faster the turnaround, you know, the more projects you can actually even complete. So.

Dario (37:27)
I don’t see why not. I don’t see why not.

Kyrill (37:48)
There is that benefit to it as well, but I don’t know. Like it’s only like a couple of weeks in since it was announced. We’ll see when it launches what applications people will start using it for. And then we can do a revisit episode and see like how it’s in like a year or two and see how it’s transformed or enhanced our industry, how we’ve potentially started using it. But who knows? It’s too early. That’s what I mean. It’s too early still.

Dario (38:10)
It’s too early, it’s too early right now. Yeah, as soon as they launch it, it’s in the red team testing phase or whatever, but as soon as it’s launched, it’ll still be too early. It’s a couple years out. It’s not around the corner. It’s a couple years out. It’s still gonna be, to fine tune it and make sure it’s good. There’s also the possibility it just never gets to be perfect. Like if you look at…

Kyrill (38:33)
Yeah.

Dario (38:34)
Mid Journey for example, they just can’t figure out hands. They’re on version 6 now and for the life of them they just can’t figure it out.

Kyrill (38:41)
It’s just like, it’s like 50 like hands is like, well, it looks like there’s two fingers in reality.

Dario (38:45)
Nah, it’s more like they can’t, they don’t understand how many fingers hands have for some weird reason. Getting them to do five fingers is like the hardest thing. So who knows, it might just get to the point where it never gets to that stage.

Kyrill (38:52)
Yep.

It’s like Bruce Almighty. It’s like Bruce Almighty with like a few fingers.

Dario (39:01)
Yeah, let’s see. I mean, again, we’re already taking advantage of AI on our end. Like we mentioned, audio editing. Obviously we use Chad GPT. I’ve been testing out Gemini lately because Chad GPT has been annoying the hell out of me. I hate the way it just answers stuff and how like, oh, you know what though? Here’s a potential downside of Sora. So Chad GPT during the golden era, you know, pre March, 2023.

just give you stuff, right? Just tell it or just do it. It was amazing. And now like you really, it wants you to do the work. It really wants you to like take the outlines and actually write the blog post yourself, which I refuse to do because I’m too busy, right?

Kyrill (39:34)
Yeah, it was good.

You

Dario (39:49)
So it could possibly get to that stage too with video AI, or text to video AI where it’s like, at the beginning, yeah, I could kind of get it there, not there yet. Then it gets to it and then it really forces you to, I don’t even know what it would force you to do, but it would make you jump through hoops and loops to try to get what you want.

Kyrill (39:49)
Yeah.

I will tell you what, I’m calling it right now. I’ll tell you exactly what it’s gonna be. All those current videos were with very simple prompts, which is like create an astronaut in space scene and it’ll create it. What it’s gonna start asking you to do is start to describe everything. Describe what he’s wearing. How does his face look? And what is his expression? Like you’d have to be very specific with every little detail.

Dario (40:23)
But you’d have to, you’d have to. No, but you would have, yeah, but Kyril, you would have to because otherwise it’ll just do random stuff, right? You would have to be very specific about everything. You mentioned streamlining your stuff further. It will probably get to the point where it might be easier to just film it sometimes, like unless you’re doing something creative like that, like someone on the moon or whatever.

Kyrill (40:34)
That’s true. Yeah.

Yeah, right?

Dario (40:51)
For basic stuff, it might just be easier to just film it. You know that classic phrase, don’t do it in post? It could get to that stage too, with text to video AI where you’re like, you know what, let’s just film it, because it might be too difficult to get it to be exactly like this. But one thing they could possibly do, and this might be a USP for us in the future, is we could like…

Kyrill (40:51)
Yeah, or animation work.

Dario (41:17)
I’m thinking, I think stable diffusion does is you can create your own like custom preset or I don’t know what the term is for it. Again, I’m not the most technical guy out there. But like imagine for a specific client, you would create their own custom AI presets for like the video side of things, right? So it has all their branding guidelines in there. No, yeah, seriously. And then even like the look for their videos, it’s all part of that.

Kyrill (41:36)
Yeah. Like Brandon guidelines.

Dario (41:47)
Cause I’m thinking if you’re doing text to video and you’re doing not just one off clips, there would have to be a consistency in the content. So you would have to have like a specific, I don’t know, preset designated to consistently get the same content out of that, out of those series of prompts, right? So that could be potentially one thing that we could use as a, as a USP. And that.

Kyrill (41:57)
Yeah, we have to be.

Dario (42:17)
that we could tell clients it’s like you have your own AI, custom AI for this project. So we will create your own profile and it’ll be everything that we create will be tailored for your content. And then that might justify the potential cost increase as well, right? Cause it has to be like custom to that client, you know?

Kyrill (42:41)
Yeah. Well, it’s the creative input that you’re putting in that you’re going to be charging for. You know, sometimes you’re not, it’s not always just pure labor that you’re charging for. It’s also what, what value you’re giving the client at the same time. There’s always that kind of little bit of a balancing act. You have to kind of determine when you are quoting for things. Um, one other.

Dario (42:59)
Not even creativity, not even creative input. It’s actually, it would just be consistency.

Kyrill (43:05)
Yeah, but you’re creating the initial creative for it to be able to be consistent, right? That’s the idea.

Dario (43:06)
Because imagine you’re going to.

Yeah, no, but the consistency would probably be the key because yeah, you could go to Sora and get something and then you might go to another, um, text to video software and get something else. But trying to match those two would be like trying to match like different camera footage together in a way, right? Or even between shot to shot might be the same situation. Like imagine if, imagine in a video, like you’re creating, let’s say 30 % of it would be AI generated, right?

Kyrill (43:30)
Yeah.

Dario (43:41)
But like what would the consistency be like imagine it like the difficulty and trying to get it to all be similar like try doing that in mid -journey try to create like a consistent Series of images and like craft a story out of that. It’s actually tough. It’s very tough So it’ll probably be the same thing with the video so that would probably be a USP and you know what? We would just be we wouldn’t really be losing money on projects. We just be shifting budget So again if production cheapens out a little bit, well that would probably just go to pre -production or post

Kyrill (43:50)
Yeah.

So it is.

Dario (44:11)
So maybe like if you were very production side heavy, it would just end up being more pre -production heavy now, cost -wise or post -production side.

Kyrill (44:22)
Yeah, well, it’s essentially kind of like how you shift money around when you’re doing animation videos, right? There you spend all the time that would have been in production more so in the pre production side of things. Like in, when you’re creating an animation video, it’s like 80 % pre -production, 20 % actual creation of the content. Because at that point, everything has been decided and then it’s just minor tweaking, right? So it’ll, it’ll be something very similar to that where it’s going to be more, more, uh, less production focused.

more pre and post side of things. Another thing that we did talk about a couple of days ago was about how potentially what you could even do with this down the road is if you were to go to a certain location remotely and film the entire space and kind of like 3D map it almost and then feed that into Sora, maybe you can tell Sora it’s like, hey, here is the space for the look.

almost like a location scout, right? Say you do a location scout and you 3D map everything there and you can, yeah, poly cam.

Dario (45:24)
Polycam and you just do like a LiDAR scan of like the whole area or like the shooting area

Kyrill (45:29)
Exactly, and then you can tell Sora, hey, we wanna create this kind of sequence of scenes with these people. Can you create a mock -up storyboard video or film of it, right? This is kind of like where it goes back to where it can help in the pre -production where you map out almost the entire scene of what it’ll look like and it’ll look very realistic. And then you can show it to a client, it’s like, hey, this is what we’re gonna go for with your people and your talent.

Let’s let’s plan for it. Like it could be very interesting to that point where even if the video is so good You may not even need to actually film it, you know that scan will Will essentially do it for you. But I don’t know that again everything we’re talking about in this video It’s all speculation of what could happen, you know based on what we’ve seen in the past It’s still early days You know, we’ll see what happens in a couple years time how it gets integrated in it, you know I I mean with the introduction of AI photos I

I still think that there is a need for photographers now. So I don’t know. We’ll have to see.

Dario (46:37)
Speaking of photographers, I was speaking with a photographer friend of mine, he mentioned that there’s people using AI to create headshots out of their photos. Have you seen that?

Kyrill (46:47)
What? No, I haven’t seen that. That’s funny. Oh man. Yeah, the…

Dario (46:48)
Yeah, yeah, yeah.

but it’s still not there. So anyways, we’re obviously gonna advocate for our side more so than the AI side. But I don’t know. I think if you’re actually immersed in it, you’ll understand the limitations. But again, who knows? They might just figure it out. Who knows? It’ll be interesting. Why did this have to happen to our generation? We’re dealing with so much stuff. COVID, this, 2008.

Kyrill (46:59)
advocate for us being there.

Yeah.

I mean, I think people also…

This is every generation that says something like this. I’m sure the previous generation of photographers and videographers were like, I know they were dealing with film and then the introduction of digital all of a sudden almost killed them. A lot more people started coming in and taking their work. They’re like, oh, why did our generation?

Dario (47:30)
They had film for the longest time, Kyrill, what are you talking about? They were dealing with film for like a hundred years.

I’m saying they at least got to they got to enjoy it. We got we got digital. We started like what mid mid to early to mid 2000s that started right and now it’s already changing. So anyways, who knows who knows what the future is going to be like.

Kyrill (47:58)
Yeah, I’m curious, but I think the successful people in our industry will learn how to adapt and integrate it in a certain way or navigate it. I think there’s still also going to be the need for clients wanting to have a good experience working with people at the same time. You know, if it’s just always going to be just one person in a room, just dealing with a computer.

Dario (48:23)
I think the budget is gonna be the determining factor for that one, Kirill.

Kyrill (48:29)
I don’t know, I’m just saying. It’s like what people will want to go for, but I don’t know. I don’t know. We’ll see. Time will tell. That’s all she wrote.

Dario (48:37)
Yeah, anyways, all right. Well, let us know if you had differing opinions or if you’re agreed with us. Let us know in the comments and yeah, we’ll catch you on the next one.

Kyrill (48:47)
All right, thanks for listening. Take care, guys.

Dario Nouri (48:51)
Thanks for listening to this episode of Creatives Grab Coffee. Please make sure to follow and engage with us on Instagram, TikTok, LinkedIn, YouTube, and your favorite podcast app. Creatives Grab Coffee is created by Laps Productions, a video production company based in Toronto, Canada. Creatives Grab Coffee is also sponsored by. My name is Mehran, welcome to Canada Film Equipment.

Thanks for listening and we’ll see you on the next one.

Sora AI and the Future of Video Production | Creatives Grab Coffee 56

Summary

Video Podcast: Sora AI and the Future of Video Production | Creatives Grab Coffee 56

Audio Podcast: Sora AI and the Future of Video Production | Creatives Grab Coffee 56

Takeaways

Timestamps

Transcript

Social Media

Related Posts