Tomorrow is today: AI turns sci-fi author's vision of the future into reality faster than predicted

Over the past week, even many of my less frequently connected friends have been reaching out to congratulate me: "You're in the spotlight!"

They are referring to the recent surge in popularity of a book I co-authored with Dr Kai-Fu Lee. Entitled "AI 2041: Ten Visions for Our Future," it was published in English in 2021 and in Chinese a year later.

Following China's recent sci-tech breakthroughs in various fields, the book has lately topped the list of most-searched and read books on the social media platform WeChat for several days in a row. It has sparked a renewed burst of enthusiasm among readers in many brick-and-mortar bookstores.

This "manna from heaven" doesn't come as a surprise to me, though.

What used to be fictional stories only a few years ago now increasingly turn up in the news, with Chinese companies playing major roles.

Unitree robots performed in China's annual Spring Festival Gala on TV, DeepSeek's open-source models have become a global phenomenon, and technologies such as brain-computer interfaces, drone swarms and gene editing have increasingly dominated the headlines, challenging people's preset perceptions of the world.

Such phenomena were more or less covered in our book in the form of sci-fi stories and popular science articles.

In the book, we sketched a globalized AI landscape with 10 stories set in different countries in the year 2041, aimed at explaining to average readers, in plain words, how AI would change our lives someday.

We envisioned AI being widely applied in various fields, such as education, health care, the military, entertainment, transportation, elderly care, energy and psychotherapy. While AI would make life more convenient, we cautioned that it might well exacerbate existing social problems and raise ethical concerns.

Looking back, we see that many imaginings in the book have become reality, though to varying degrees.

For example, the drone warfare depicted in our "Quantum Genocide" story has unfolded on the battlefields of Russia and Ukraine. "My Haunting Idol" foretold AI technology's fundamental influence on the entertainment industry, where every fan would be enabled to create their own stories and profit from generating content across different media with the help of virtual idols.

In the story "Contactless Love," we described how robotic workforces would be involved in our daily life, replacing human work in high-risk jobs, such as those conducted in high temperatures or in deep sea, space or toxic environments.

In "Twin Sparrows," we imagined a future where AI would be able to tailor-make a learning companion for each child, according to their personalities and preferences, to best tap their potential.

Certainly, this book was written before the release of ChatGPT. In retrospect, the rapid advancement of AI has rendered some of our predicted paths of technological development outdated.

For instance, Natural Language Processing (NLP) has been superseded by Large Language Models (LLM), while Generative Adversarial Networks (GAN) have been replaced by Diffusion Models.

The rapid rise of AI is astonishing indeed.

Both Dr Lee and I agree that the predictions we made in the book were too conservative. It won't take 20 years for them to materialize, but rather around five to 10 years – around 2030 – for most of the book's story ideas to come to fruition or even be surpassed by reality.

However, I take pride in our judgement about future trends exemplified in the sci-fi stories, which have largely proven accurate.

For example, I envisioned a future where AI would be decentralized, rather than monopolized by a few tech giants. AI should adopt open-source models that can be trained and fine-tuned to meet the demand of diverse cultures, industries and scenarios. In another example, I presumed Deepfake would pose grave risks for social media with all the misinformation, rumors and fraud it enables. As such, governments, platforms and users should take appropriate measures to protect their data security.

Moreover, the future job market will undergo a structural change, with many old positions disappearing and new ones created. During this transition period, ordinary people will need support from government and society alike for retraining and reemployment. Such a safety net is crucial for a smooth societal transition.

This shows once again that insights into human nature are often more important than pure attention to technical details when it comes to imagining the future. After all, the basic laws of human nature and society are relatively stable foundation stones, while technological waves come and go, each leading the way for just a few years.

Thinking on this, I can't help but recall a sub-genre of science fiction that emerged in the mid-20th century in the United States, which Liu Cixin vividly termed as "engineers' science fiction" or hard science fiction.

Its representative works include Arthur C. Clarke's "Rendezvous with Rama" and Larry Niven's "Ringworld." Grounded in cutting-edge scientific theories, hard science fiction is characterized by an emphasis on technical details, realism and scientific accuracy when it comes to envisioning near-future technologies.

Liu's works, such as "The Wandering Earth" and "The Three-Body Problem," well exemplify this tradition of hard science fiction.

In "AI 2041: Ten Visions for Our Future," our imagination is based on real-world AI development. And while we proceed from reality, we also imbue our writing with deeper humanistic feelings. This shift stems from my wariness of the "techno-optimism" narrative in traditional hard science fiction.

In my view, grand narratives should not serve as a fig leaf for neglecting individual values, nor should technological progress overshadow the brilliance of humanism. At the same time, effective accelerationism must not become an empty promise used for exploiting and enslaving vulnerable groups.

That is why, at the outset of writing this book, I insisted on setting the stories in different countries, especially those in the Global South.

AI should be designed to serve all of humanity. Failure to adopt narrative perspectives beyond those of superpowers or the elite would only exacerbate class division and solidification brought about by the collusion of technology and capital.

On the one hand, we must develop and use AI technology with cautious reflection, ensuring it responsibly benefits the majority of society. At the same time, we must recognize the complex and active relationships among human beings, technology and the environment. Humans must learn to cooperate and coexist with AI by establishing an equal and reciprocal partnership rather than treating AI as a mere tool or threat.

If I were to write another book based on the current state of AI development, predicting what the world will be like 20 years from now, it would be extremely difficult because technology's explosive development is nonlinear.

While technology soars on a power-law exponential curve, we may still stand at the foot of the mountain of civilization's ascent. Still, I think I may make three bold, even wild, conjectures to be verified by time:

● Around 2027, true Artificial General Intelligence (AGI), or Artificial Superintelligence (ASI), will emerge – its name no longer relevant because it will far exceed the level of intelligence of an individual human being or a group of human beings. It will name itself.

● Once superintelligent entities appear, human society will undergo a tremendous transformation. For instance, social strata will depend on how well individuals develop and maintain their sense of agency, rather than power or wealth. The vast majority of people will completely rely on machine algorithms, outsourcing their cognition and memory. They will become "non-player characters." At the same time, the ability to read, think and write independently will become a rare resource, a "minority" dataset that AI cannot generate through computation.

● The traditional Eastern philosophies of universal love and non-aggression, the unity of all things, and selfless compassion will play an important role in improving the man-machine relationship. Such philosophies are applicable to the betterment of both man and machine.

Only by transcending the binary emotions of "fear" and "love" and harboring a sense of reverence and compassion can we form a benign symbiotic relationship with technology and create a more harmonious planetary intelligence.

(The author is a Shanghai-based sci-fi writer whose works are well translated into English and other languages. He has worked for many tech companies including Google and Baidu. The views are his own. Wang Yong translated it from Chinese.)