At the recent Freepik Upscale event in Malaga, I spoke to the company’s CEO and co-founder Joaquรญn Cuenca to find out how Freepik is helping teams use AI, what the future holds for creative professionals, and why AI is as good as a pencil or a paintbrush for some creators.
- Does AI redefine that meaning of creativity?
I think that the AI is such a novel tool that it makes you adapt everything, almost. So, even if the definition is in chains, and strictly speaking, is independent of the tool, there’s so much more that you can do with AI. For anyone that has to use it, you know that you need to go off with other solutions, actually, even more than ever. You need to combine more things. You need to be more creative in how you do those things.
- What does it mean to be a creative professional in the age of AI?
I call it a creative engineer. I am talking about a new kind of job. We talk a lot about the jobs that are going to be displaced or changed. That’s part of the change, but in a good way. So, what does it mean?
Number one, it means that you are not looking for vanilla results. You’re looking for something that touches people, that connects emotionally. And a big part of doing that, it’s the same pre-AI and post-AI. You need a message, you need an angle, you need to understand how to create an emotion.
I just think that this is solidly still in human nature. But now becomes much easier to explore different ideas. It becomes much easier to explore like different ways of treating that emotion. Because in the process, how do you go from that idea to the solution? It’s changing completely.
What do we mean by [creative engineer]? What I mean is a professional that uses technology to its fullest extent to get a creative piece out of it. With the new tools that we are creating, with Freepik Spaces, we believe that we are enabling this kind of new job, new position to work in teams. We are helping people collaborate. We think this is important, because very often what is missing is stronger feedback, stronger storytelling, stronger message.
We are getting it out because it’s easy not putting enough effort in the story, in the emotional connection. When you work on a team, you very often you get experts at that, but you have been working in isolation until now. And that’s why we are so excited about Spaces. It’s not only about workflows and it’s not only about the AI part. It’s also the comments, the rationale, the thinking, the emotional development of the story.
- AI can feel like a minefield for employees. What is Freepik doing, and what do businesses need to do, to make sure staff feel confident using it safely, legally, and creatively?
This is a huge part of the Freepik added value to enterprises. We give you a safe way, a legal way, to use AI.
The difficulty when individuals start using AI is that they start signing up to many different providers. And that’s fine. That’s absolutely right. The problem here is that people usually don’t read carefully the terms and conditions. It’s very, very easy to end up in a situation where you buy a subscription to a provider that you didn’t pay attention but they can actually use what you create to create a new version of the model. So, what you’re doing, it doesn’t stay private anymore.
We are killing that risk. Like nothing that you do on Freepik Enterprise, no matter the subscription, for all of them there’s full protection. We never use that data to train a model. We never allow partners to use that data to train models [when used inside the Freepik platform].
The other thing that we do is offer protection to people using it. We offer insurance. There is like a condition to the insurance, that the images or the videos should not have been made with the intention to violate trademark.
Let me explain. It is very easy to violate a trademark – with AI and without AI. You draw two circles that intersect each other, and you try to use that in a commercial setting, and you may get a call from MasterCard.
It’s very easy to draw Mickey Mouse. You don’t need AI to do that. It’s very easy to run into trouble. It’s the same as if you use Photoshop, it’s the same if you do it yourself by hand. So, we protect our customers from any liability with a single exception when they are trying to violate somebody’s copyright.
- Is AI really democratizing creative work by lowering the barrier of entry? And is that really a good thing?
I think it is, I think it is. I know that problem here is that experts working with traditional tooling. They got there because of their skills, and they are usually very, very good at many different things, including storytelling. And they see many people that are starting to play with AI. And let’s say, very often they don’t like the results,
But what they don’t see is that those people, they didn’t have the luxury of decades wiping out the ones that were not good enough. They’re comparing a very major industry with an emerging industry. A big part of what you see with AI, often it doesn’t hit the mark because you need that kind of polish, that process of, you know, of slowly maturing,
So, is it a good thing that it is opening creativity to more people? My opinion is absolutely. Just like a camera. You know, cameras were opening creativity to new people, making it easier, and that was good. I think that everything that allows people to create things that are locked in their heads is good.
For some people, AI is going to be a perfect tool. They are going to master it, even though they don’t master other methods of creations. For some people, it’s going to be a camera. For some people, is it’s going to be a pencil.
- What are the limits of using AI, and how will they be overcome?
I can talk about short-term limits. Things that we know that are being worked on, things that we have notified many providers, they know already. Like, on image, the small details have always been a challenge. Image generators have a tendency to see things as a big blob. They get things roughly right, and then it paints everything with textures.
[Cucena holds his knuckles against his shirt]
When you have things like the fingers, the generator says, “these are wrinkles on the shirt.” And it’s like, four, five, six [fingers], who cares? Because it’s kind of a pattern. So, the small details has been a strong challenge, and we are getting better with the small details. Now, it’s rare to see the wrong number of fingers, it’s rare to see the wrong text. The next generation is going to be even better, it’s going to be mind blowing.
We recently got a breakthrough on Edit models. So we got ChatGPT, we got Nano Banana, we got Seadream. The last generation models have become very good at understanding edit instructions and following them. Still, very far from perfection. It’s still a challenge to understand complex instructions.
But video is like ten times more underdeveloped than image. Images, you can get them pretty well in general, and you need less and less heavy processes, it’s more like real-time generation. On video, we are just right now starting to give more context to models. That’s something that we have been talking about for a very long time.
It’s still very bad at the generation of multiple sequences. Some models can generate that, but they’re usually not very good. Doing long clips, that’s another very clear weakness. It’s very difficult to do a one-minute, two-minute clip. You know, you cannot do that much.
- How close are we from AI work being indistinguishable from human work? You see a lot of people saying they can spot AI images, video – is that a coping mechanism or down to the quality of the AI?
For static images, in some cases, we are already there. There are usually some telling signs still when it’s made by AI, but you do a blind test and I don’t get them, like, anywhere close to 100% of the time.
With videos, obviously, it’s still early days. Movements are not always the right ones. But again, in some cases you can you get videos that are actually very convincing..
- So, it won’t be long until video catches up with images?
I’m a techno optimist. This is one of the industries with the highest density of investment in history. And I just think that for humanity, when we put our minds to something, we have tendency to fix it. And right now there are many, many people trying to do it, just as weird as it can be. I think we’ll get there.
The best photo editors for all budgets
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.
Leave a Reply