Original source: Geek Park
Author|Tong Qi, Hu Xiaomeng
![]()
Image source: Generated by Unbounded AI
Since the Industrial Revolution, human beings have gradually become accustomed to getting along with machines day and night for hundreds of years, from the initial panic to the current ignorance, and now the media has become an extension of the body, the mobile phone has become the integration of the media, and the machine has even become a part of the human body. The relationship between humans and machines is becoming more and more intimate, and machines are no longer just objects and dead things. Many people hope that machines will have real intelligence and emotion, and then have the ability to think independently and be autonomous, and have a deep emotional exchange with humans.
![]()
The development of large artificial intelligence (AI) models represented by ChatGPT not only heralds the acceleration of the era of General Artificial Intelligence (AGI), but also greatly increases the possibility of turning human imagination into reality. As the application of AI in daily production and life has greatly improved efficiency and convenience, generative AI is now triggering a new round of excitement and fear: when AI lives and works with us, or when AI becomes a member of our city, the feeling of being surrounded by ubiquitous and amorphous technology fuels a complex emotion of self-awareness and fear of machine intelligence.
When technology has long been integrated into our daily life, becoming ubiquitous and amorphous, affecting people’s thinking, perception and behavior like a ghost, perhaps the daily life of artificial intelligence is not a new topic, but we are unconsciously entering the era of symbiosis with AI. With the wave of large models represented by ChatGPT and the emergence of generative AI, the commercialization path of AI has been further accelerated, making it have a non-negligible impact on the daily work, consumption, and entertainment of the general public. This means a revolutionary change in the way of human-computer interaction: visual interaction represented by graphical user interfaces has developed to the extreme, and has begun to turn to access all scenes and interact with humans through natural language interfaces. Behind this new media technology revolution are not only business, political and social issues, but also ethical, privacy and impartiality issues, but also the fundamental shaking of humanism that may occur under the influence of artificial intelligence and algorithmic decision-making.
One: When artificial intelligence becomes the technical background of daily life
When AI becomes a part of everyday life, humans usually don’t feel the presence of technology, which is the “background relationship” of the four types of human-technology relationships proposed by Don Ihde. Contextual relationship [human → (AI/world)] means that AI as a technological background, such as electricity and the Internet, has become a part of daily life, and people are often only aware of its existence when the technology fails. Of course, in daily life, there are three other kinds of relationships between humans and artificial intelligence: embodied relationship [(human-AI)→world)], that is, AI as a functional extension of human beings, to recognize the world and practice; interpretive relationship [human→ (AI-world)], that is, human cognition of the world, is representational, transformed or interpreted by AI, for example, the recommendation algorithm influences and shapes human cognitive patterns to a certain extent, and it is heterogeneous [human→AI-(world)] , focusing on the autonomy of technology, that is, when AI becomes the object of cognition, the world presented through AI becomes a technological artifact.
So what are the moments in the daily lives of contemporary people that are influenced by AI? This question can now also be answered through dialogue with artificial intelligence. When asked about “a scenario in which a modern person interacts with AI unconsciously in a day”, ChatGPT’s answer is quite comprehensive (see Table 1), and it even reminds us that “everyone’s lifestyle habits and needs may be the same, so the actual situation may be different.” ”
In other words, natural language interaction tools powered by generative AI technology can already tell us that AI is becoming a part of work, leisure, socializing, and even sleep, and its technological potential covers all aspects of a person’s life. In this way, from the perspective of a larger living space, the city where humans live has become a city where humans and machines coexist, and perhaps in the future, it will become a city where humans and artificial intelligence coexist. This means that AI also “lives” in the city, and thus becomes an indispensable member of the city. From ChatGPT’s answers, a possible future has been revealed, that is, human beings have inevitably entered a state of life that is deeply integrated with artificial intelligence.
However, in the face of the tide of artificial intelligence technology, there are two primary risks that need to be vigilant when dealing with the relationship between people and technology: one is the failure of the technology, and the other is the loss of control of the technology.
![]()
First, as Don Ede argues, technology that penetrates into everyday life is only recognized when it fails. Once the technology fails, it will bring immeasurable harm to society. After the “7.20 Zhengzhou Heavy Rainstorm” in 2021, a popular commentary article “Zhengzhou after the disaster: When a city suddenly lost the Internet” recorded the failure of all habitual lifestyles such as online shopping, electronic payment, urban transportation, and information dissemination after the heavy rain disaster. In other words, the normal operation of human society in the information age depends on the superimposed algorithms, electricity, networks and other background technologies in human life as infrastructure and prerequisites. In the era of artificial intelligence, it may be difficult for human society to get rid of the risk of technological failure.
Second, human life is intertwined with artificial intelligence, but will artificial intelligence be free from human control and operate on its own? Recently, a research team from Stanford University in the United States released a very interesting study, they made a sandbox game with a large language model, in which 25 generative agents (Generative Agents) “live freely” in a virtual environment: these agents are like humans, waking up, making breakfast, and going to work in an interactive sandbox environment every day; React to this and adjust your plans when appropriate, resulting in unexpectedly complex behaviors (such as throwing a Valentine’s Day party). Users can interact with these 25 agents using natural language. The research team noted that this work reveals the architecture and interaction patterns that enable the simulation of credible human behavior by fusing large language models with computer interaction agents.
When artificial intelligence can interact freely in a simulated environment, it also indicates to some extent that artificial intelligence may be out of human control in the near future. It is not difficult to imagine that with the development of technology, artificial intelligence and robots may achieve a deeper integration, and when this happens, can human beings still become the dominant ones in life? Or, more specifically, can human beings continue to dominate the order of the city? At this time, what human life will rely on to maintain the basic operation, and what kind of values should be used as the benchmark for common life, will become an important question that cannot be avoided under the impetus of the new technological revolution. In this sense, the so-called value alignment of AI with human values is particularly important.
2: When AI enters the era of large models
The starting point of modern society, or globalization, was the Age of Discovery at the end of the 15th century, when the advancement of knowledge and technology, and the development of navigation technology, European explorers were able to explore and conquer new continents, and the global ecology and geopolitics were turned upside down. It can be said that the Age of Discovery mainly realized the expansion of technology into space, and the modern capitalist mode of production and the modern life of urbanization are closely related to this.
The development of contemporary information technology has realized the expansion and crazy appropriation of time. Our time and brains are filled with the massive amount of information carried and provided by information technology, and generative AI continues to accelerate this information overload. But the day on Earth is still 24 hours, and this contradiction has turned our lives into an interaction between “social time” and “digital time”. In the future, artificial intelligence will become an intermediary, or continue to evolve into a “compressor” of time and civilization. From machine computing to artificial intelligence, breakthrough technologies continue to emerge, and human civilization has entered the “era of deep technology”.
Large models are undoubtedly the most important development trend of AI technology, and they are also the basis for AI industrialization and daily life. With the explosion of OpenAI’s GPT large model, global technology companies and scientific research institutes are vying to be the first to invest in the competition of large models, and artificial intelligence has now entered the era of large models. Large AI models, also known as pre-training models and foundation models, are models trained based on massive amounts of data with a large number of parameters, which can be adapted to a wide range of downstream tasks. Based on the ideas of transfer learning and the latest advances in the field of deep learning, these models demonstrate surprising “emergence” capabilities with the help of large-scale computer systems and significantly improve the performance of various downstream tasks.
The large model marks a paradigm change in the development of AI technology, and digital and intelligent systems in various fields will be built on the large model in the future. Entering the era of large models also means the arrival of the stage of AI industrialization development, which provides a path for AI to achieve standardization, modularization, and automation. Taking the AI intelligent customer service that people often come into contact with as an example, in the past, the application of AI intelligent customer service in various fields such as healthcare, e-commerce, and finance required redesign and R&D training every time, and its algorithm model could not be reused. Therefore, the high cost of technology research and development and labor costs limit the large-scale application of AI intelligent customer service. The large model does not need to reconstruct different scene models for application scenarios, nor does it need to start training algorithm models again, but only needs to quickly extract and generate scenario-based, customized, and personalized small models on the basis of large models, so as to realize industrial assembly line deployment in different industries, vertical fields, and functional scenarios, and have the advantages of on-demand use and high efficiency and economy.
It can be seen that the large model is becoming the future intelligent technology infrastructure, reconstructing the existing business model, and forming a “model-as-a-service” technology and industrial ecology in the future. In the future, digital business will be divided into large-model infrastructure enterprises, small-scale model application enterprises in vertical industries, and model applications and services that are more suitable for individual users. The establishment and development of this ecosystem will more widely empower applications in various industries, accelerate the digital transformation and intelligent development of all fields of society, and improve the production efficiency of the whole society.
The accumulation and integration of technologies such as large models, multimodal, and generative algorithms represented by GPT have given birth to the explosion of generative artificial intelligence (AIGC), and what the general public perceives is mainly at the level of “model as a service”, that is, text, pictures, audio, video and other content generation services for public users. It can be said that as a new type of content production and dissemination, generative AI can bring a new round of production and consumption revolution, and take the lead in media, e-commerce, games, film and television and other industries with a high degree of digitalization and high demand for content.
![]()
Timeline of Artificial Intelligence Development (drawn by Hu Xiaomeng)
Three: When ChatGPT Starts the Language Interaction Revolution
The creation of the so-called general artificial intelligence is considered to be an important reason for the revolutionary change of artificial intelligence, and versatility has also become an important criterion for revolutionary technological change. In fact, it wasn’t until the advent of computers that the versatility of information technology became a possibility. Artificial intelligence is also evolving towards this standard. Xiao Yanghua, a professor at Fudan University, pointed out that all technological breakthroughs have been led by humans, which is itself the result of intelligence, but general artificial intelligence is a technology that aims to approach or even surpass human intelligence, and it is a technological change about intelligence itself, and its importance and danger are far greater than before.
When users use generative AI represented by ChatGPT, they can input a piece of human language text into the user interface, which can be a complete sentence or part of a sentence, which is used to prompt the large model to generate the next part of the text. This prompt text is the prompt word (). How to generate content that is both credible and creative becomes a task that requires constant testing and tweaking. The process of continuous prompting is the process of human interaction with the large model through language, the process of value exchange, and the process of continuous training of the personal version of GPT, which finally makes it a tool with more personal attributes and more in line with personal values.
![]()
Natural language processing is regarded as the “crown jewel” of artificial intelligence, and allowing machines to understand human language has always been the direction of countless artificial intelligence scientists and researchers. The advent of language models represented by ChatGPT has announced that language is no longer the privilege of human civilization, and machines can also induct, summarize, apply, and give examples of human language, and output it through human language (instead of the machine language represented by 0 and 1). The human-computer interaction mode has also expanded from graphical interaction interface to natural language interaction interface.
Language is creative, but at the same time deceptive. The power of GPT has led most users to perceive the content generated by GPT as trustworthy and usable. Human trust in artificial intelligence even exceeds trust in their own judgment, which makes artificial intelligence (especially the most important application direction of artificial intelligence - algorithm recommendation) become a certain authoritative existence. Although many scholars have mentioned the potential harm caused by the authority of algorithms, with the advancement of technology, the authority of technology will become higher and higher in society. Our daily lives are not only computable, but also required to be calculated at high speeds, and as a result, we often have to spend huge amounts of energy and computing power in exchange for time to adapt to the accelerated era brought about by accelerated computing. The question is, artificial intelligence is not always intelligent, algorithms are not always reliable, and even in some critical scenarios, will our lives be forced to be in the hands of algorithms?
Similarly, whether humans will have intimate feelings with artificial intelligence through verbal communication has become a topic that is constantly being discussed. In fact, falling in love with an AI-powered virtual lover is no longer a shocking thing, and technology companies are constantly developing corresponding products for users to choose from. Historian and philosopher Yuval Noah Harari argues that while there is no evidence that AI is conscious or sentient, the fact that AI can make people feel emotionally connected to them is enough to establish an intimate relationship between humans and machines — an unusual intimacy that may change our worldview. In addition, he made a very thought-provoking conclusion: because of the amazing ability to manipulate and generate language, AI has invaded the “operating system” of human civilization.
4: Continue to explore the value of people in the era of human-machine symbiosis
"In the age of algorithms, the pie is so big that most people can only get a small piece. "The lives and rights of ordinary people are increasingly being ceded to technology, but the dividends of technological development will still be highly concentrated in the hands of a few. In the era of ChatGPT, artificial intelligence has become a part of productivity, and it is foreseeable that our learning mode, thinking mode, question and answer mode, and behavior mode will be affected by a large model similar to GPT. All of these scenarios will be built or reconstructed based on the AI model. In other words, the way we perceive the world, the way we interact with machines, will change radically.
Former U.S. Secretary of State Henry Kissinger believes that ChatGPT heralds a knowledge revolution in which humans need to think about and respond to the far-reaching impact of artificial intelligence on the way humans exist and know, and determine the human-machine relationship and the moral and strategic role of humans. This reminder can be said to be sobering. Living in the digital age, interacting with artificial intelligence and following the advice of algorithms is, to some extent, delivering subjectivity to algorithms, which, no matter how intelligent, are only looking for connections between things. Individuals increasingly believe in the results given by algorithms, and their lives are calculated, even calculated. This involves the continuous interaction between autonomous decision-making, human subjectivity, and algorithmic decision-making and the machine intelligence behind it.
There is no doubt that we are at an inflection point in the technological era: when artificial intelligence becomes the technical background of daily life, when AI technology enters the era of large models, when ChatGPT opens up the transformation of language interaction… The technological revolution may be another great wave in the field of life. To some extent, however, all philosophers are reminding us not to gently enter the age of symbiosis with AI, and not to be overly seduced by technological optimism and developmentalism. When the superiority of human intelligence is replaced by artificial intelligence, when humans rely on machines and algorithms to determine every aspect of their daily lives, where is the uniqueness and future of human beings? The intricacies and multiplicity of the body, mind, emotion, and behavior are what it means to be human. In the era when humans and AI live together, how to continue to reflect on the relationship between humans and machines, between people, and how to continue to protect human dignity must be included in the discussion and consideration at this moment.
