Artificial intelligence has moved beyond simple calculations and information searches into the realm of human creation. It draws pictures and composes music in seconds, suggests ideas, and writes sentences. Generating results in an instant that once required humans decades of honing skills is no longer surprising.

The game industry is no different. Going beyond how to utilize AI in the development process, we must now talk about an era where human-made content and AI-generated content compete in the same market. If even "creativity," once thought to be an exclusively human domain, is being caught up to by machines, a single question naturally arises.

"Are we cooked?"

This is the question posed on stage at gamescom dev by Fawzi Mesmar, Creative Director at Ubisoft. Mesmar is a veteran developer with over 23 years of game development experience, having worked at King and DICE before currently serving as Creative Director for Beyond Good and Evil 2 at Ubisoft. He has taught game design and creativity, and in 2024 published the book 'Demystifying Creativity: On Originality in Game Development', which explores creativity and originality.

At gamescom dev held in Cologne, Germany on the 26th, Mesmar gave a presentation titled 'Creativity in the Age of AI: Practicing Creative Sobriety'. Instead of asking whether AI will replace human creativity, the session broke human creativity down into several elements, compared them with AI, and explored what competitive edge humans can still retain.

And his answer was neither a simple "yes" nor "no."

In chess, art, and ideas… in certain areas, humans have already lost

There is no need to turn a blind eye to what AI does better than humans. Mesmar cited 'Creative Proficiency' as the first element of creativity. This refers to the mechanical and technical skills needed to express oneself, such as in music, dance, or art. Robots can already move faster than humans and play complex music that is difficult for humans to perform. In chess and Go, AI surpassed top human players long ago.

So, are humans finished in this regard? Mesmar's answer was "Yes and no." While we may give ground to machines in technical proficiency, what people actually want to see is human victory and failure, and the struggle that unfolds in the process. Not many people want to spend hours watching two AIs play a perfect game of chess. When looking at competition and proficiency, the focus of our interest remains human experience.

AI also has clear advantages in problem solving. Mesmar cited German psychologist Karl Duncker's famous 'Candle Problem' as an example. When given a candle, matches, and a box of thumbtacks, and asked to attach the candle to a wall so wax does not drip on the table, many people fail to view the 'box' containing the tacks as a tool to solve the problem. This is due to 'functional fixedness,' seeing objects only in terms of their intended purpose.

AI is relatively free from such fixedness. If an element exists within the data, AI can quickly find things that humans miss due to their own biases. However, there are limits here as well. Defining what the problem is and choosing which data to use to solve it must still be decided by humans. At this point, Mesmar sees room for humans to stay one step ahead.

The situation is similar in creative work. Humans utilize 'analogical thinking' to find solutions in completely different fields. Just as Nike designer Tinker Hatfield was inspired by the exposed structure of the Centre Pompidou in Paris to design the 'Air Max' with visible internal air cushioning, new ideas are born when seemingly unrelated experiences connect. 'The Last of Us' similarly expanded the concept of Cordyceps fungi infecting insects and controlling their behavior into a fungus that infects humans.

However, AI now makes even these connections exceptionally fast. Mesmar explained that when he asked AI for an idea that could cause a real-world zombie outbreak, it immediately suggested a fungal parasite, and when asked for natural elements to create zombies, it even came up with ideas like a 'coral zombie' connecting coral and humans. Making logical connections across several steps within existing data is an area where AI already excels.

There is data AI has not trained on: 'myself'

The concept Mesmar introduced here is also the title of his talk: 'Creative Sobriety.'

Inside a person's mind, there exists an associative network where numerous concepts and experiences are interconnected. For instance, the word "color" might bring to mind "white," which leads to "milk," milk to "cow," cow to "horse," and horse to "unicorn." The closer the connection between concepts, the easier they are to think of; the further the connection, the greater the likelihood of generating a more original idea.

Mesmar described what anyone can easily think of at the first connection as 'Inevitable Ideas.' Inevitable ideas are not inherently bad; if anyone can think of it, it is an idea that can be brought to market that much faster. The problem is that in the AI era, it becomes difficult for humans to compete in this domain.

"If your idea is 'purple' and there is no purple in the world yet, you must start making it right now. Because even at this very moment, there are probably five other people making purple in their garages."

AI can generate these inevitable ideas at a speed beyond comparison with humans. Creative sobriety as defined by Mesmar begins with consciously observing which step of association your idea came from. Originality is ultimately **statistically, an idea that others are unlikely to think of**, and understanding where you are allows you to consciously go further.

So how can we go further than AI? Mesmar's answer is surprisingly simple: start from data that AI cannot learn from.

That data is one's own life.

Everyone has their own experiences, memories, and perspectives on the world. What events you experienced, what emotions they left you with, and why they moved you cannot be entirely identical to anyone else's. Mesmar shared an example from his own life, when he was kayaking in New Zealand, got surrounded by killer whales, and escaped with the help of a local drug dealer. Connections born from such experiences exist exclusively within one's own associative network.

To illustrate this, he also shared a failure story involving a friend. After watching Netflix's 'Squid Game', the friend thought making a game based on it would earn a lot of money. The friend continued development by reproducing the games featured in the show as they were, but by the time it was completed, thousands of 'Squid Game' knockoffs had already flooded the market.

Mesmar pointed out that the issue was that his friend watched 'Squid Game' and immediately jumped to 'making Squid Game.' If millions of people watched the same show, countless game developers among them likely reached the same connection. Instead, one should ask themselves, "Why did this work move me?" and "How does it connect to parts of my own life?"

"Instead of recreating what the world showed us, we must create based on how the world moved us. Recreating what we directly see is something AI can do very easily."

Ultimately, the AI era demands a higher level of originality from humans than before. In the past, even ideas anyone could think of were competitive enough if brought to market quickly, but once AI begins producing them at an overwhelming speed, humans cannot compete on pace.

"That is why we are not cooked yet. Rather, we are now under pressure to be more original than ever before. We must be more honest about who we are, and create something from data that AI can never train on."

AI can create, but it cannot take 'creative responsibility'

Mesmar went a step further here. Creativity matters to humans not solely for generating ideas superior to AI, but because human creations spark thoughts and works in others, sometimes influencing the far future.

The figure he cited as an example was Albert Einstein. Developing an interest in unseen forces after looking at a compass in his childhood, Einstein studied mathematics and physics, and exposed himself to numerous cutting-edge ideas while working at the patent office. Yet his research was not born of existing knowledge alone. He imagined what the world would look like if he rode a beam of light at light speed—adding his imagination to humanity's accumulated knowledge.

The same applies to gaming. Mentioning the North American video game crash of 1983, Mesmar explained that while the industry was driven to the brink of collapse, games created by Nintendo and developers like Shigeru Miyamoto exerted a profound influence on the industry ever since. And he did not stop at comparing Einstein to Miyamoto.

"You might think I just compared Miyamoto to Einstein. But I will go one step further: I am comparing both of them to you."

What he sought to address was the universality of creation and the accompanying 'Creative Responsibility.' What a creator puts out into the world can impact unexpected people and become the seed for them to create something new. No one knows how a single game you build today might inspire someone decades down the line.

Mesmar likened this to "planting seeds for trees whose shade we may never sit under." Pouring the society, beliefs, and experiences of this present moment into a creation to leave behind for the world, leaving open the possibility that someone will one day discover it and build something else, is the responsibility held by a creator.

And this was the clearest line he drew between AI and humans.

"AI cannot own its voice. It cannot be censored, taken to court, or directly defend the lived experiences embedded in the art it creates."

AI can produce output, but it does not have a life to explain why it made it. Humans, on the other hand, embed their experiences, beliefs, and the era they inhabit into their work, and take responsibility for the outcomes of releasing it to the world.

Hence, Mesmar views AI as a 'tool of expression,' no different from Photoshop, game engines, or musical instruments. What matters is not what the tool can produce, but what the person using it seeks to express. Creators must ask themselves: Which of my experiences will I infuse into this work? What is impossible for others to replicate? And am I willing to be honest and vulnerable enough to put a piece of myself out into the world to be judged?

The volume of content produced by AI is likely to expand beyond human reach. Mesmar did not deny this reality. However, he viewed this less as an 'existential threat' to human creators and more as a challenge.

"I want you to embrace your creative responsibility. Be honest with yourselves, and stay true to the era you live in and the things that move you. The work you create should reflect who you are, your perspective, and how the world has impacted you."

That was the ultimate meaning of 'Creative Sobriety' as articulated by Mesmar. It is not about making more or building faster than AI, but about being conscious of where your ideas come from, breaking away from conventional connections, and delving deeper into your own experiences.

Returning to the initial question: in the midst of AI advancement, are we cooked?

Mesmar's answer ultimately comes close to "not yet." However, in an era where machines churn out obvious ideas far faster than humans ever could, humans must become more human than before.

Not recreating the world AI has trained on, but creating based on how that world moved 'you.' That was his conclusion on how human creations will survive amid the flood of AI-generated content.