2026/09/14

看见迹象 Seeing Signs

 


少焉,月出于富丽敦酒店( Fullerton Hotel)之上,徘徊于新加坡河天之间。

走出新加坡亚洲文明博物馆,清风徐来,暑气未消。本来热闹喧腾的室外灯光秀已经结束。抬眼望见左前方天上一轮明月,想起今天是农历七月十六日,正是苏东坡《赤壁赋》里写的七月既望啊。

准备了两个月,多方研究,今天终于交出还算满意的成绩单,完成导览《汇流:卢浮宫的莫卧儿、萨法维与奥斯曼艺术珍品》展览。我把这场九十分钟的导览题名为“看见迹象:伊斯兰美术的转译与本土化”。

迹象的概念来自《古兰经》第41章第53节:我将把我的种种迹象显示在四方,并显示在他们自身中,直到他们明白这是真理。阿拉伯语 āyah(复数 āyāt)既指天地与生命中可供辨识的迹象,也用来称《古兰经》的经文。

“迹象”和文图学讲的文本(Text)在观看方法上相通,但目的略有出入。日月星辰运行和人的自身,都是真主的迹象,引导人明瞭创造的秩序与启示的真理——迹象召唤人们去体察世界。文本则是意义生成、转译与理解的场域。万物皆文本,文字、图像、器物、声音、气味、身体、空间和光线,通过媒介、历史、文化语境与观看者相互联系,观点也随时间、地点和解读方式发生变化。文图学由此追问:意义怎样形成?怎样传播?怎样被转译?又怎样在不同地方获得新的生命?

导览伊始,我朝文图学会会员听众们的手心逐一喷洒液体,有人搓了搓,马上闻出:这是玫瑰露。是的,展品中有物件图像,是和玫瑰露有关。

虽然导览的主题集中于历史学家形容的三大火药帝国Gunpowder Empires),时代在公元16-18世纪,不过,这些文明区域早在7世纪,甚至更早,就和中国有商贸和文化往来,“物证”就是亚洲文明博物馆展示的黑石号沉船出水文物。我带着大家浏览了河南巩县的唐代青花盘、巨型执壶,留下视觉印象。随后,我们看见在印度莫卧儿帝国的珍珠贝母执壶盘组合,以及类似造型的金属执壶盘组合,可知器型的延续关系。

景德镇的青花瓷出口到萨法维帝国(今伊朗),深受喜爱,中国的吉祥瑞兽例如麒麟也传播到异域,出现了外观类似,内里别样的新形式。这种承接和转换在文化交流,尤其是文物艺术品的模仿制作过程中,屡见不鲜。我找了明代的麒麟纹盘和现场类似的展品对照,大家会心一笑,频频点头说:有趣!

这两个相隔两百年的麒麟纹盘属于同一套视觉语法:青白色彩、同心圆构图、中央瑞兽和外围缠枝花卉。明代的牡丹缠枝石草图像比伊朗的放射状如菊花和丰肥的叶片细致具体;盘中央的明代麒麟是独角、龙首、鳞身,被火焰围绕,仿佛奔腾着向祥云昂首回颈。没有看展签之前,我还以为伊朗的这件蓝地留白大盘画的是狮子呢,可爱!

看展的乐趣之一,是亲眼认识或发现微妙的细节。曾经听连文思老师介绍过刻有波斯诗句的烛台,听她娓娓道来,十分感动。我端详那些貌似花纹的古波斯文字,注意到不只顶端有,底座也有,顶端的文字是诗句;底座则是拥有者的题记。我把诗句翻译成:

我的心因爱恋美人而燃烧,肝肠如焚;
爱情每一刻,又以新的烙印灼烧着我。
我与烛火相对,恰如飞蛾;
若再向前一步,我的双翼便会燃烧。

好奇这样的诗句,如果搭配音乐和波斯语,听起来是怎样的感觉呢?于是我借用 AI语音,录制波斯语朗读,配上伊朗音乐家凯汉·卡尔霍尔(Kayhan Kalhor)演奏的卡曼切琴(kamancheh)。琴声低回悠远,诗句沿着器身周旋,热烈的爱火点亮了烛台。

导览最后,来到奥斯曼咖啡馆的图像和杯盘。我播放了土耳其音乐家库德西·埃尔居内尔(Kudsi Ergüner)演奏的古典音乐。那幽幽的内伊笛(ney),穿越咖啡馆里读诗、下棋的人们,满是生活的迹象。

 

2026912日,新加坡《联合早报》上善若水专栏

 

Seeing Signs

I Lo-fen

Before long, the moon rose above the Fullerton Hotel and lingered between the sky and the Singapore River.

As I walked out of the Asian Civilisations Museum in Singapore, a gentle breeze began to blow, though the heat of the day had not yet faded. The lively outdoor light show had already ended. Looking up, I saw a bright full moon hanging in the sky to my left. Then I remembered that it was the sixteenth day of the seventh lunar month—the very “day after the full moon in the seventh month” described by Su Shi in his Former Ode on the Red Cliffs.

After two months of preparation and research from many different angles, I finally felt that I had produced a reasonably satisfying result today, completing my guided tour of the exhibition Crosscurrents: Masterpieces of Mughal, Safavid and Ottoman Art from the Musée du Louvre. I entitled this ninety-minute tour “Seeing Signs: Translation and Localisation in Islamic Art.”

The concept of “signs” comes from verse 53 of Chapter 41 of the Quran: “We shall show them Our signs in the horizons and within themselves until it becomes clear to them that it is the truth.” The Arabic word āyah (plural āyāt) refers both to “signs” in the heavens, the earth, and living beings that invite recognition, and to individual verses of the Quran.

As a mode of seeing, the idea of “signs” has much in common with the concept of “text” in Text and Image Studies, although their purposes differ somewhat. The movements of the sun, moon, and stars, as well as the human self, are all signs of God, guiding people toward an understanding of the order of creation and the truth of revelation. “Signs” call upon us to observe and contemplate the world. “Text,” by contrast, is a site where meaning is generated, translated, and interpreted. Everything can be read as text: words, images, objects, sounds, smells, bodies, spaces, and light are interconnected through media, history, cultural context, and the viewer, while perspectives change with time, place, and modes of interpretation. Text and Image Studies therefore asks: How is meaning formed? How is it transmitted? How is it translated? And how does it acquire new life in different places?

At the beginning of the tour, I sprayed a liquid into the palms of the members of the Text and Image Studies Society who had joined me. Some rubbed their hands together and immediately recognised the fragrance: rosewater. Yes—among the objects and images in the exhibition were works associated with rosewater.

Although the tour focused on what historians have called the three great “Gunpowder Empires,” mainly from the sixteenth to the eighteenth centuries, these regions had already maintained commercial and cultural contacts with China by the seventh century, and even earlier. The material evidence can be found in the objects recovered from the Tang Shipwreck displayed at the Asian Civilisations Museum. I first led the group past Tang-dynasty blue-and-white dishes from the Gongxian kilns in Henan and a monumental ewer, leaving them with a visual impression. We then encountered a Mughal ewer-and-basin set made of mother-of-pearl, as well as a metal ewer-and-basin set of similar form, allowing us to see how vessel types continued across regions and periods.

Blue-and-white porcelain from Jingdezhen was exported to the Safavid Empire, in present-day Iran, where it was greatly admired. Auspicious Chinese creatures such as the qilin also travelled abroad, giving rise to new forms that looked familiar on the surface yet were transformed within. Such processes of inheritance and adaptation are frequently encountered in cultural exchange, especially in the imitation and remaking of works of art and material culture. I found a Ming-dynasty dish decorated with a qilin and compared it with a similar object in the exhibition. Everyone smiled knowingly and nodded repeatedly: “Interesting!”

The two qilin dishes, made two centuries apart, belong to the same visual grammar: a blue-and-white palette, concentric composition, an auspicious beast at the centre, and scrolling floral ornament around the rim. The Ming-dynasty dish renders peonies, scrolling stems, and rocks with greater specificity and delicacy than the Iranian work, whose ornament is more radial, with chrysanthemum-like blossoms and fuller leaves. At the centre of the Ming dish, the qilin has a single horn, a dragon-like head, and a scaled body surrounded by flames. It appears to be galloping forward while turning its head upward toward the auspicious clouds. Before reading the label, I had thought the large Iranian dish, with its reserve decoration against a blue ground, depicted a lion. How charming!

One of the pleasures of visiting an exhibition is being able to recognise or discover subtle details with one’s own eyes. I had once listened to Dr Lien Wen Sze introduce a candlestick inscribed with Persian poetry, and I was deeply moved by the way she unfolded its story. As I examined what at first looked like decorative patterns, I realised that the Persian inscriptions appeared not only at the top but also around the base. The upper inscription consists of poetry, while the lower one records an owner’s inscription.

I translated the poem as follows:

My heart burns with love for the beloved; my very entrails are aflame.

At every moment, love sears me again with a fresh brand.

I face the candle flame like a moth;

one step closer, and my wings will burn.

I became curious: what would these lines sound like in Persian, accompanied by music? So I used an AI voice to create a Persian recitation and paired it with Iranian musician Kayhan Kalhor playing the kamancheh. The instrument’s low, lingering sound seemed to circle around the candlestick together with the verses, while the ardour of love itself appeared to ignite the flame. Kalhor is internationally known as a virtuoso of the kamancheh, the Persian spike fiddle.

At the end of the tour, we arrived at images and tableware associated with an Ottoman coffeehouse. I played classical music performed by Turkish musician Kudsi Ergüner. The haunting sound of the ney seemed to pass through the figures in the coffeehouse as they recited poetry and played chess, filling the scene with the signs of everyday life.

12 September 2026, “Shang Shan Ruo Shui” column, Lianhe Zaobao, Singapore.

 

2026/08/29

AI双语新加坡 AI Bilingual Singapore

 


“Bilingual”这个词,对新加坡人并不陌生。

几十年来,双语教育不仅是一项教育政策,也塑造了新加坡人的生活经验:英语让人进入全球网络,母语维系文化、家庭与身份。一个人能够在两种语言之间转换,也意味着能够在不同知识系统和文化语境之间移动。

到了AI时代,新加坡开始谈一种新的双语

2026年,新加坡推出全国人工智能影响计划National AI Impact Programme),提出未来三年协助一万家企业深化AI应用,同时支持十万名工作者成为 AI Bilingual所谓“AI双语,不是会说两种自然语言——母语和英语,而是既熟悉自己的专业领域,又能够理解和运用AI,把AI真正带进工作流程。首批重点包括会计、法律等非科技专业,随后还将扩展到其他领域。

数码发展及新闻部长杨莉明今年谈到“AI Bilinguals”时特别指出,这个比喻之所以容易被新加坡人理解,正因为多数新加坡人都是接受双语教育。她把一个人的专业知识比作自己的母语,而AI则像需要重新学习的另一种语言。

这个比喻比必须学习AI”有意思得多,也非常新加坡派

学一门语言,从来不只是背几个单词;真正的双语,也不是知道两套词汇,而是知道在什么时候说什么、怎样理解对方、怎样转换语境。同样地,AI双语也不等于会打开ChatGPT、写几个提示词。一个律师必须先懂法律,一个教师必须先懂教育,一个医生必须先懂医学;AI能力的意义,是让原有的专业判断获得新的工具,而不是让工具取代专业判断。

新加坡为什么此刻如此强调这种能力?数字让我们一目了然。

2023年,新加坡只有4.2%的中小企业采用AI;到了2024年,这一比例已经升到14.5%,一年增长超过三倍。大型企业同期也从44%升至62.5%。与此同时,2024年新加坡数字经济规模达到1281亿新元,占GDP18.6%AI显然正在从少数科技公司的实验室,进入越来越普通的企业。

更值得注意的是人。新加坡资讯通信媒体发展局IMDA调查显示,73.8%的受访工作者已经在工作中使用AI;其中58%用于构思与头脑风暴,54%用于写作与编辑。AI最先进入的,恰恰不是科幻电影里的机器人岗位,而是我们每天都在做的思考、写作、沟通和整理。

这也解释了为什么“AI双语“AI专家更值得推动

新加坡不是要求十万人都变成程序员,而是希望十万人在自己的专业里学会AI”。杨莉明举过一个很好的例子:投资经理可以利用AI检验投资判断,但最后作决定的仍然是人。她特别强调,AI应用的目的不是让人替AI工作,而是让AI为人工作。

人力部的数据也说明,目前AI带来的主要变化不是大规模裁员。2026年的调查中,28.5%的企业已经采用AI;其中只有6.2%报告减少人手或招聘,反而有18.9%重新设计工作内容,13.9%创造新的AI相关岗位,70.7%表示员工生产力提高。换句话说,眼下更明显的不是职业消失,而是职业正在重新书写自己的语法。

语言会改变人的思考方式,工具也会。

过去,新加坡的双语教育训练人们在不同语言和文化之间转换;今天,“AI双语又提出一种新的转换能力:在人的专业知识与机器能力之间往返,在效率与判断之间保持距离,在接受AI帮助的同时,仍然知道什么时候应该怀疑它。

所以,我觉得“AI双语真正值得认可的,不是这个新名词有多时髦,而是它重新思考个老问题:一个人要用什么语言和技能,才能记忆以往,面对未来的世界?

在新加坡,答案是:“AI双语

 

2026829日新加坡《联合早报》“上善若水”专栏

 

AI Bilingual Singapore

I Lo-fen

The word “bilingual” is hardly unfamiliar to Singaporeans.

For decades, bilingual education has been more than an educational policy; it has also shaped the lived experience of Singaporeans. English connects people to global networks, while their Mother Tongue sustains culture, family ties, and identity. To be able to move between two languages is also to move between different systems of knowledge and cultural contexts.

In the age of AI, Singapore has begun speaking of a new kind of “bilingualism.”

In 2026, Singapore launched the National AI Impact Programme (NAIIP), which aims over the next three years to help 10,000 enterprises deepen their adoption of AI, while supporting 100,000 workers to become “AI Bilingual.” An “AI Bilingual” is not someone who speaks two natural languages—English and a Mother Tongue—but someone who is both well versed in a professional domain and able to understand and apply AI, integrating it meaningfully into actual workflows. The initial focus is on non-tech professions such as accountancy and law, with other fields to follow.

When Minister for Digital Development and Information Josephine Teo spoke this year about “AI Bilinguals,” she pointed out that the metaphor is intuitively easy for Singaporeans to understand precisely because most Singaporeans are products of bilingual education. She compared a person’s domain expertise to one language and AI to another language that must be learned. 

The metaphor is far more interesting than simply saying that “everyone must learn AI.” It is also very Singaporean.

Learning a language has never been merely a matter of memorising a few words. True bilingualism does not mean knowing two sets of vocabulary; it means knowing what to say, when to say it, how to understand the other party, and how to move between contexts. In the same way, being AI bilingual does not simply mean knowing how to open ChatGPT and write a few prompts. A lawyer must first understand law, a teacher must first understand education, and a doctor must first understand medicine. The value of AI capability lies in giving professional judgement new tools, rather than allowing tools to replace professional judgement.

Why is Singapore placing such emphasis on this capability at this particular moment? The numbers make the answer clear.

In 2023, only 4.2 per cent of Singapore’s small and medium-sized enterprises had adopted AI. By 2024, the figure had risen to 14.5 per cent—more than tripling in a single year. Among larger enterprises, adoption rose over the same period from 44 per cent to 62.5 per cent. Meanwhile, Singapore’s digital economy reached S$128.1 billion in 2024, accounting for 18.6 per cent of GDP. AI is clearly moving out of the laboratories of a small number of technology companies and into increasingly ordinary businesses.

Even more striking is what is happening among workers. A survey by Singapore’s Infocomm Media Development Authority (IMDA) found that 73.8 per cent of surveyed working individuals were already using AI tools at work. Among these AI users, 58 per cent used AI for brainstorming and ideation, while 54 per cent used it for writing and editing. The first areas AI has entered are not the robotic jobs imagined in science-fiction films, but precisely the thinking, writing, communicating, and organising that we do every day.

This also explains why promoting “AI Bilinguals” matters more than trying to turn everyone into an “AI expert.”

Singapore is not asking 100,000 people to become programmers. It wants 100,000 people to learn how to “speak AI” within their own professions. Josephine Teo has offered a useful example: a portfolio manager can use AI to sharpen an investment thesis and stress-test positions, but the final decision remains with the human professional. She has also emphasised that AI implementation is not simply about getting humans to make AI work; it is about getting AI to work in support of humans.

Data from the Ministry of Manpower likewise suggest that the main effect of AI at present is not large-scale job displacement. In its 2026 survey, 28.5 per cent of firms had begun adopting AI. Only 6.2 per cent of AI-adopting firms reported reductions in headcount, while 18.9 per cent had redesigned roles, 13.9 per cent had created new AI-related jobs, and 70.7 per cent reported improvements in worker productivity. In other words, what is most visible at the moment is not the disappearance of occupations, but occupations rewriting their own grammar.

Language changes the way people think. Tools do too.

In the past, Singapore’s bilingual education trained people to move between different languages and cultures. Today, “AI bilingualism” proposes another kind of ability to switch: moving back and forth between human professional knowledge and machine capability, maintaining a distance between efficiency and judgement, and knowing when to question AI even while accepting its assistance.

That is why, to me, what is truly worth recognising about “AI Bilingual” is not how fashionable the new term sounds, but the way it reopens an old question: What languages and skills must a person possess in order to remember the past and face the world of the future?

In Singapore, the answer is:AI Bilingual.


29 August 2026, “Shang Shan Ruo Shui” column, Lianhe Zaobao, Singapore.

 

2026/08/15

乾隆宸翰何真假 How Genuine Is Qianlong’s Imperial Calligraphy?

 


乾隆宸翰何真假

衣若芬

 

走进博物馆,看见一幅古画,上面题着诗,旁边钤满大大小小的印章。许多人第一时间会问:这是不是乾隆皇帝亲笔?

这个问题看似简单,却未必容易回答。

前些时候在台北故宫看展览,面对一件传为乾隆御题的书画作品,我忽然想到:如果今天有人工智能能够百分之百模仿乾隆的笔迹,我们还能一眼分辨真假吗?而如果不能,那么我们究竟在寻找什么样的

乾隆皇帝是中国历史上留下题跋最多的皇帝。有人开玩笑,说他是盖章狂魔,或者是弹幕达人。然而,关于乾隆题字的真假问题,却一直存在。

故宫收藏中,有些作品上的御题被认为出自乾隆亲笔;有些则可能由宫廷朝臣代笔。对于书法鉴定家来说,这当然是重要课题。用笔提顿习惯是否一致?线条是否自然?结体章法是否符合乾隆书风?这些都成为判断的依据。

可是,当我站在展柜前凝视那些文字时,却觉得问题似乎不止于此。

即使某一段题字不是乾隆亲手所写,它是否就因此失去价值?

在现代人的观念里,作品往往与作者紧密绑定。我们推崇原创,重视亲笔,相信独一无二的创作主体。因此,一旦发现作品出自代笔,便容易产生失落感。然而,在传统中国宫廷文化中,事情未必如此。

一件御题作品的形成,往往涉及复杂的生产过程。构思诗文、书写题跋、装裱制作、内府收藏……最终呈现在我们面前的,并不是单一个人的劳动成就,而是一整套宫廷机制共同运作的结果。

换言之,所谓乾隆宸翰,有时候代表的不只是乾隆的手笔,而是乾隆的意志。从这个角度来看,真假便出现了新的层次。

真,不只是笔迹的真。

真,也是制度的真、时代的真、文化语境的真。

我们以为自己在观看文物,其实也在观看一套关于文物的叙事。

从文图学的角度来看,这一点尤其明显。

一件书画作品从来不是单纯的图像。画面、题跋、印章、装裱、收藏记录、展览说明,共同构成一个可供阅读的文本系统。观众所理解的真迹,往往不是来自某一个笔画,而是来自这一整套文图关系。

因此,当我们讨论乾隆宸翰真假时,实际上讨论的已经不只是书法鉴定,而是关于真实性的文化想像。

有趣的是,这个问题在人工智能时代显得格外尖锐。

今天的生成式人工智能能够模仿画风、模仿文体,甚至模仿历史人物的书法。如果输入足够多的乾隆书迹样本,也许未来真的能够生成一幅几可乱真的乾隆御笔

到那时,真假又该如何界定?

如果一首诗是人工智能根据《御制诗集》学习后写出的;一幅字是根据乾隆书风生成的,那么它究竟是假作,还是另一种形式的再创造?

这个问题听起来很现代,却意外地又让我想到清代宫廷。

因为所谓原创作者的观念,本来就是近代以来逐渐形成的。乾隆时代的大量文化生产,早已建立在人与人协作的机制之上。今天的人机协作创作,不过是在新的技术条件下重新提出相似的问题。

于是,我忽然觉得自己寻找的答案已经改变了。

我不再急着判断那几个字是不是乾隆亲手所写。我更感兴趣的是:为什么我们如此在意它是否亲笔?

也许,真正值得思考的,不是乾隆宸翰究竟是真是假,而是我们如何定义真实。真实可以是一段笔墨留下的痕迹,也可以是一种时代精神的延续;可以是一位作者的亲手书写,也可以是一套文化机制的共同创造。

当人工智能不断挑战原创与作者的边界时,乾隆留下的那些题诗与印章,仿佛从历史深处发出回响——它提醒我们,真实性从来不是单纯的技术判断,而是一种文化选择。

因此,乾隆宸翰何真假?

或许答案并不在光学分析仪器下,也不在专家鉴定报告里。

答案更可能藏在我们观看它的目光之中。因为每一个时代,都在重新定义什么是真,什么是假;而每一次定义,也都映照着那个时代对于历史、作者与自我的理解。

2026815日,新加坡《联合早报》“上善若水”专栏


How Genuine Is Qianlong’s Imperial Calligraphy?
I Lo-fen

Walk into a museum and encounter an old painting inscribed with a poem and covered with seals of all sizes. Many people’s first question is: Was this really written by the Qianlong Emperor himself?

The question sounds simple, but it is not necessarily easy to answer.

Some time ago, while visiting an exhibition at the National Palace Museum in Taipei, I stood before a painting and calligraphy work said to bear an imperial inscription by Qianlong. A thought suddenly occurred to me: if artificial intelligence today could reproduce Qianlong’s handwriting with one hundred percent accuracy, would we still be able to distinguish the genuine from the fake at a glance? And if not, what kind of “authenticity” are we actually looking for?

The Qianlong Emperor left more inscriptions and colophons than any other emperor in Chinese history. People sometimes joke that he was a “seal-stamping fanatic” or a master of “bullet comments.” Yet questions about the authenticity of Qianlong’s inscriptions have persisted for a long time.

Among works in palace museum collections, some imperial inscriptions are considered to have been written by Qianlong himself, while others may have been executed on his behalf by court officials. For connoisseurs of calligraphy, this is naturally an important issue. Are the habits of brush pressure, lifting, and turning consistent? Are the lines natural? Do the structure of individual characters and the overall composition conform to Qianlong’s calligraphic style? All of these become criteria for judgement.

Yet as I stood before the display case and gazed at those words, I felt that the question went beyond this.

Even if a particular inscription was not written by Qianlong’s own hand, does it therefore lose its value?

In the modern understanding of art, a work is often closely bound to its author. We prize originality, value autograph works, and believe in the uniqueness of the creative individual. Once we discover that a work was written by someone else on the purported author’s behalf, disappointment easily follows. Yet things were not necessarily understood in this way within the culture of the traditional Chinese court.

The making of an imperially inscribed work often involved a complex process: composing the poem or text, writing the inscription, mounting and producing the finished object, and incorporating it into the imperial collection. What finally appears before us is therefore not the achievement of a single individual, but the result of an entire court system working together.

In other words, “Qianlong’s imperial calligraphy” sometimes represents not only Qianlong’s hand, but Qianlong’s will. Seen from this perspective, the distinction between genuine and fake acquires new layers.

Authenticity is not only the authenticity of handwriting.

It is also the authenticity of an institution, an era, and a cultural context.

We think we are looking at an artefact, but in fact we are also looking at a narrative constructed around that artefact.

From the perspective of Text and Image Studies, this becomes especially evident.

A work of painting and calligraphy is never simply an image. The painting, inscriptions, seals, mounting, collection records, and exhibition labels together form a textual system that can be read. What viewers understand as an “authentic work” often derives not from a single brushstroke, but from this entire network of relationships between text and image.

Thus, when we discuss whether Qianlong’s imperial calligraphy is genuine or fake, we are in fact discussing more than the authentication of calligraphy. We are discussing a cultural imagination of authenticity.

What makes the question especially interesting is that it becomes particularly acute in the age of artificial intelligence.

Today, generative artificial intelligence can imitate painting styles, literary styles, and even the calligraphy of historical figures. If supplied with enough samples of Qianlong’s handwriting, perhaps one day it really will be able to generate a “Qianlong imperial inscription” almost indistinguishable from the real thing.

How, then, should we define what is genuine and what is fake?

If a poem is written by artificial intelligence after learning from the Imperially Composed Poems, and a piece of calligraphy is generated according to Qianlong’s style, is it a forgery, or another form of re-creation?

The question sounds thoroughly modern, yet it unexpectedly brings me back to the Qing imperial court.

The very idea of the “original author,” after all, took shape gradually in the modern era. Much of the cultural production of Qianlong’s time was already based on mechanisms of human collaboration. Today’s human–AI collaborative creation merely raises similar questions again under new technological conditions.

At that moment, I realised that the answer I had been seeking had changed.

I was no longer in a hurry to determine whether those characters had been written by Qianlong’s own hand. What interested me more was this: why do we care so deeply whether they were?

Perhaps what truly deserves our attention is not whether Qianlong’s imperial calligraphy is ultimately genuine or fake, but how we define “authenticity.” Authenticity may be the trace left by brush and ink; it may also be the continuation of the spirit of an age. It may be writing executed by the author’s own hand; it may also be something collectively created by an entire cultural system.

As artificial intelligence continues to challenge the boundaries of originality and authorship, the poems and seals left by Qianlong seem to echo from deep within history. They remind us that authenticity has never been merely a technical judgement, but a cultural choice.

So, how genuine is Qianlong’s imperial calligraphy?

Perhaps the answer does not lie beneath an optical analyser, nor in an expert authentication report.

More likely, it is hidden in the way we look at it. Every age redefines what is true and what is false; and every such definition reflects that age’s understanding of history, authorship, and itself.

15 August 2026, “Shang Shan Ruo Shui” column, Lianhe Zaobao, Singapore.