A few years ago, AlphaGo made its debut and turned the Go world upside down when it defeated Lee Sedol. 🌪️
For any field to remain solid, it needs experts who’ve poured time and effort into it, and they must be respected, forming a pyramid-like structure. 🏛️
In the case of Go, it started as playful tossing of stones on the floor or table, but it took a long journey to reach where it is today. Experts honed their skills through repeated matches, while observant individuals without the expertise generalized their insights into rules. Masters of framework creation used these materials to build educational systems, allowing many children to grow up dreaming of becoming Go champions. All these efforts aim for the vision of becoming a ‘master who can defeat any opponent in Go.’ Someone becomes the next Lee Sedol, reaching the pinnacle and astonishing the world, driving the entire industry forward. 🚀
But then, AI threw down the gauntlet at this stronghold. 🏰 It mercilessly split the tower of Babel in two. Machines barged into the human game and shredded it. Since then, AI has become so robust that it’s near impossible for humans to win against it. In online Go games, the number of people cheating using AI is growing, and the quality of moves is judged by whether they match AI’s placements rather than analyzing the players’ strategies. Human moves are no longer respected, and people have started studying AI’s moves, which lack meaning. As Go became harder to play relying solely on human skills, people gradually lost interest, and thus, new Go academies have all but vanished. 📉
It’s been said for a while that simple repetitive tasks would be replaced by AI. This is already quite advanced, but fields requiring human creativity were thought to be harder for AI to tackle. Can we call Go a creative field? While human moves are creative, the mechanism operates under certain rules, making it hard to see as a true adversary to algorithms. Perhaps it wasn’t really an art form. However, AI these days is finding immense joy in creation. 🎨
Remember the buzz around NLP (Natural Language Processing) after OpenAI’s release of GPT-3 in 2020? The following year, the text-to-image AI called DALL-E was introduced. Then, in October, the more advanced model Disco Diffusion was announced. Around this time, artists began using these AI tools. Soon after, MidJourney was introduced, making it even easier for the general public to create AI images. In April, OpenAI released DALL-E2, kicking off a new era in AI text-to-image. 🌟
This artwork was created using MidJourney on Discord with the prompt ‘a gloomy girl in a white skirt.’ Can we really say I ‘drew’ it? It’s a bit ironic that the artists best suited for creating stunning images with these tools are actually writers. The current model generates images based on prompts, much like finding the most effective search term in a search engine. While there may be differences in the artistry of prompts, the idea of finding that magic keyword to get what you want is the same. 🔍✨
Businesses with clear, explainable rules are easier to algorithmize. But when infinite reference data and irregular combinations are needed, the story changes. The more complex and irregular the process of prioritizing and applying suitable rules, the more incomprehensibly complex the algorithm becomes. Deep learning solves this complexity with immense data, but its biggest drawback, except for a few models, is the inability to explain meanings or reasons clearly. Thus, deep learning hasn’t yet made a major impact in business industries where precise information is crucial. However, there’s one area where precision is less critical—the realm of emotions, the art world. 🎨❤️
As mentioned, art was thought to be the hardest area for AI to invade due to its need for creativity. But I see it differently. Given the evolution of deep learning models, the field where AI can make the biggest impact now is likely the arts. There’s no need for logical explanations or accountability. Some degree of imperfection can be pushed as artistic or original. With existing creative techniques and abundant references, the market’s increasing focus on reconstruction or tweaks of existing outputs sends positive signals for AI’s creative activities. Machines lack creativity and originality, but AI will overcome this with persistence and rapid repetition. GPUs are becoming more efficient and affordable. 💡💪
MidJourney already produces quality work that rivals general illustrators within seconds. Not satisfied? Slightly tweak the prompt or press a button for a redo. AI will offer a new suggestion in seconds without a single complaint. 🤖💫
Artificial intelligence is one of the most profound things humanity is working on. More profound than fire or electricity.Sundar Pichai (2020), CEO of Google
That doesn’t mean that producers in the art world will be replaced by AI with a simple flip of a switch. But, it does mean that we need to consider how to navigate these changes. It feels like facing an unpredictable adversary. Is this constantly evolving technology really the master key humanity needs in these challenging times? 🔑🤔
Leave a Reply