Amidst the feast of AI poetry and painting, have humans become mere spectators? When machines learn to create, what is the unique value of humans?

Author/Liu Lan

Produced by/Xinzhai Business Review

Reflection in the era of large models: Are we really making progress?

As AI emerges prominently in the field of creation at an astonishing speed, should humans re-examine their position and value? In today's rapid development of AI technology, have we lost ourselves in the fog of technology and forgotten our original intentions?

When AI initiates a new era of intelligent creation in the name of poetry and painting, are humans ready to embrace this unprecedented challenge and opportunity? Under ChatGPT's brushstrokes, art is no longer exclusive to humans but a vast world where intelligence and humans dance together.

Is this the joy of technological progress hidden behind, or the repositioning of the human role?

I. AI poetry and painting are popular across the internet, but humans are still trapped in trivial daily routines?

In 2023, with the emergence of ChatGPT, the field of AI creation ushered in an unprecedented revolution. This 'chat expert in the AI world' is not only adept at conversation but also demonstrates astonishing talent in painting and writing. People only need to tap lightly on the keyboard and engage in an imaginative dialogue with ChatGPT, and lifelike paintings of various styles will appear before their eyes.

This is not only a technological miracle but also a warm moment when dreams come true.

When AI shines brightly in the field of creation, humans seem to be deeply confused. Originally, we expected robots to help us share household chores, giving us more time to pursue poetry and distant places.

In reality, robots start writing poems and painting, while humans are still busy with trivialities of life.

This seemingly humorous joke actually implies a profound reflection on the development direction of artificial intelligence: Does the rapid development of generative artificial intelligence truly align with human intentions?

The convenience of AI creation cannot be ignored. Whether it's a lonely traveler under the vast starry sky, a cup of coffee and a book on a spring afternoon, or even a fantastic scene that exists only in dreams, they can all be vividly presented in this wave of AI. With its powerful natural language processing capabilities, ChatGPT significantly lowers the threshold for creation, allowing even painting beginners to easily get started and enjoy the fun of creation.

Da A is a designer who originally needed to spend hours completing a design draft. However, since the assistance of AI painting tools, he only needs to input a few keywords to quickly obtain multiple design schemes. Although work efficiency has greatly improved, he feels an unprecedented emptiness. He finds himself increasingly reliant on AI, while his creativity gradually fades.

Behind the convenience of AI creation lies limitations that cannot be ignored.

Simply put, although AI can quickly generate high-quality content, it cannot fully understand human emotions and contexts at present. Therefore, when creating works involving deep emotions or complex contexts, AI often struggles.

Secondly, although AI creations exhibit diverse styles, they often lack a unique soul and personality. This is because AI creations are based on the analysis and imitation of massive artworks rather than true originality.

AI creations may also involve copyright issues. Since AI creations may also be protected by copyright law, this undoubtedly poses new challenges to the dissemination and use of works.

At the same time, are we restricting the creativity of AI and even humans themselves with the 'cage' constructed by data?

This prompts him to reflect: Is the efficiency we pursue really worth it?

II. The latecomer's trap and the debate over routes

In the field of AI creation, the latecomer's trap is an issue that cannot be ignored.

Taking China's 'Hundred Models War' as an example, major vendors have launched their large models, attempting to secure a place in the field of intelligent creation. Behind the myriad of models and vast parameters lies the status quo of technological convergence and lack of innovation.

Many latecomers blindly follow the footsteps of pioneers, attempting to quickly catch up by replicating their technologies and experiences. However, the result is often that pioneers reap the benefits, while latecomers find themselves trapped in endless pursuit and imitation.

The root of this latecomer's trap lies in the blind pursuit of technological development and the neglect of innovative capabilities. In the field of intelligent creation, true innovation often stems from a deep understanding of the essence of technology and profound insight into human needs, rather than simply piling up parameters and imitating pioneers.

Additionally, the debate over routes is also an important issue in the field of AI creation.

Currently, the autoregressive generative route is the hottest technical route. It relies on the 'brute force aesthetics' of big data, large models, and massive computing power, continuously brushing parameters to enhance model performance. However, is this route really worth our collective bet?

Reviewing the history of technological development, it is not difficult to find that no single route has always dominated. Choosing the 'brute force aesthetics' route requires advantages in computing power, data support, and capital investment. Do we really possess the strength to compete with pioneers in these aspects? Perhaps we should pay more attention to technological routes that can truly solve human problems and improve the quality of human life.

Amidst the upsurge in AI creation, we must also be vigilant against cognitive traps. Many believe that the ultimate goal of AI creation is to achieve AGI (Artificial General Intelligence), making computers as intelligent as humans. However, this cognition actually contains significant misconceptions.

Firstly, achieving AGI is an extremely complex process involving a deep understanding and simulation of human intelligence. Currently, our understanding of human intelligence is still very limited, so there is still a long way to go before AGI can be realized.

Secondly, even if we achieve AGI, can it truly replace humans as the main creators? Creation is not just a display of skill but also a communication and resonance of the soul. This is something that AI cannot replace.

Therefore, we should view the development of AI creation more rationally.

III. ChatGPT leads the creation revolution, have humans deviated from the original intention of intelligence?

With the release of GPT4 and the emergence of humanoid robots like Figure01, the path of large models seems to be an inevitable trend in the development of artificial intelligence.

On this seemingly bright path, many traps and challenges are hidden. Firstly, the latecomer's trap cannot be ignored, and secondly, the route trap is also a major issue.

Finally, the cognitive trap is also something we need to be vigilant about. In the pursuit of large models, have we truly understood the essence of intelligence? Are we too superstitious about the power of data and algorithms, neglecting human wisdom and creativity?

A scientist pursuing AGI gradually found himself trapped in cognitive misconceptions. He relied too heavily on data and algorithms, neglecting the complexity and diversity of the real world. Ultimately, he realized that true intelligence is not just the processing and analysis of data but the perception and understanding of the real world. Therefore, he began to adjust his research direction, placing more emphasis on interaction and connection with the real world.

When exploring the potential and value of AI creation, we cannot help but ask: What is the true direction of intelligence? Should we continue to pursue the 'brute force aesthetics' of big data, large models, and massive computing power, or explore more efficient, environmentally friendly, and user-friendly intelligent technologies?

Some believe that the path of large models is the only way to AGI. Only by continuously brushing parameters and improving performance can we achieve true intelligence.

Others hold different views. They believe that the essence of intelligence lies in the perception and understanding of the real world, rather than the simple processing and analysis of data.

Therefore, we should place more emphasis on interaction and connection with the real world, exploring intelligent technologies that better align with human wisdom and creativity.

Studies have shown that new intelligent technologies such as embodied intelligence exhibit strong potential in perception, cognition, reasoning, learning, execution, and other aspects. These technologies can not only better adapt to the complexity and diversity of the real world but also interact with humans more naturally and efficiently.

Perhaps they can provide us with intelligent solutions that better meet human needs. At the same time, we should also realize that intelligence is not static.

Looking back, we cannot help but ask: Have we lost our way in the wave of pursuing AI creation? Have we overly pursued technological breakthroughs and efficiency improvements, neglecting human wisdom and creativity? In future development, how should we find the balance between intelligence and humans?

Although AI technology is powerful, it is not omnipotent. It cannot replace human emotions and creativity, nor can it fully understand the human inner world.

Therefore, we should use AI technology as an auxiliary tool rather than a substitute.

[Conclusion]

Is the collision between ChatGPT and AI painting a technological miracle or a challenge to human creativity? In this intelligent revolution, what have we gained and what have we lost?

In this era of rapid technological development, every technological innovation seems to open a door to a new world for us. However, when AI starts to shine brightly in artistic fields such as poetry and painting, we cannot help but ask: Is this really the intelligent future we desire?